The University of Chicago Law School Roundtable

Volume 2 | Issue 2 Article 9

1-1-1995

Economics and the Problem of DivorceAnn Laquer Estin

Follow this and additional works at: http://chicagounbound.uchicago.edu/roundtable

This Article is brought to you for free and open access by Chicago Unbound. It has been accepted for inclusion in The University of Chicago LawSchool Roundtable by an authorized administrator of Chicago Unbound. For more information, please contact unbound@law.uchicago.edu.

Recommended CitationEstin, Ann Laquer (1995) "Economics and the Problem of Divorce," The University of Chicago Law School Roundtable: Vol. 2: Iss. 2,Article 9.Available at: http://chicagounbound.uchicago.edu/roundtable/vol2/iss2/9

ARTICLES

Economics and the Problem of Divorce

ANN LAQUER ESTIN

For a non-economist interested in theoretical and policy aspects offamily law, the economic theories of divorce are a puzzle. Economists, andlawyers influenced by economic analysis, tend to conceive of divorce as aproblem in efficiency. This analysis treats family law as a structure withinwhich rational individuals negotiate self-interested agreements concerningmarriage and divorce. Recent writers agree that current divorce practices arenot efficient, but they differ significantly in their, prescriptions for policyreform. Their recommendations range from abolishing judicial divorce proceed-ings entirely, and substituting a system requiring parties to reach privatedivorce agreements, to restoring the notion of fault, recast in terms of contractbreach, to judicial divorce proceedings. Although many of the specific recom-mendations are contradictory, they are nonetheless presented with a claim ofscientific authority.

Among the economists who propose dramatic changes in our presentdivorce rules is Gary Becker of the University of Chicago. Professor Becker,who was awarded the Nobel Prize for his work in 1992, is generally creditedwith inaugurating the movement of economists into the study of non-marketbehavior.' Based on the premise that all human behavior is rational, allhuman behavior becomes grist for the economic mill.2 One of ProfessorBecker's chief concerns has been an economic analysis of family life: marriage,divorce, the bearing and raising of children.3 In his academic writing, Profes-

Ann Laquer Estin is an associate professor of law at the University of Colorado. Shewould like to thank Eric Talley, Peg Brinig, and Lloyd Cohen for their comments onearlier drafts.

1. See generally Richard A. Posner, Gary Becker's Contributions to Law and Econom-ics, 22 J Legal Stud 211 (1993). On Becker's prize, see Gary S. Becker, Nobel Lecture:The Economic Way of Looking at Behavior, 101 J Pol Econ 385 (1993); Gary S. Becker,When the Wake-up Call is from the Nobel Committee, Bus Wk 20 (Nov 20, 1992).

2. See Richard A. Posner, Economic Analysis of Law (Little, Brown, 4th ed 1992).Judge Posner is largely responsible for popularizing and disseminating this approach inlaw; his "textbook-treatise," first published in 1973, is now in its fourth edition. Inaddition, he has authored dozens of other books and articles, and a growing body ofjudicial opinions, taking an economic approach to various policy issues.

3. Becker's major work on the family was first published in 1981, Gary S. Becker,A Treatise on the Family (Harvard, 1981) and later revised in 1991, Gary S. Becker, A

518 Roundtable

sor Becker has approached the family largely in the positivist tradition. In lessacademic settings, he has offered explicit policy recommendations. Among theseis his proposal that unilateral no-fault divorce should be abolished because ofits harmful effects on women with children.4 This Article assesses the applica-tions of normative economic argument to marriage and divorce. In addition toProfessor Becker's work, recent literature in this area includes a growing bodyof theoretical and empirical work by other economists,' as well as a wave oflaw and economics writing on marriage and divorce.6 After considering these

Treatise on the Family (Harvard, enlarged ed 1991) (subsequent citations are to thisenlarged edition). For an assessment of his approach, see generally Mark Blaug, TheMethodology of Economics, or, How Economists Explain 220 (Cambridge, 2d ed 1992);Isabel V. Sawhill, Economic Perspectives on the Family, 106 Daedalus 115, 120-24 (Spring1977).

4. Gary S. Becker, Finding Fault with No-Fault Divorce, Bus Wk 22 (Dec 7, 1992)(Unilateral "[n]o-fault divorce laws discourage married women from leaving the work forceto care for their young children. . . ."). See also Becker, 101 J Pol Econ at 396 (cited innote 1) ("[T]he theory does indicate that no-fault divorce hurts women with childrenwhose marriages are broken up by their husbands."). Becker's argument is discussed morefully in the text accompanying notes 119-24 and 132-37.

5. These include Kristian Bolin, The Marriage Contract and Efficient Rules forSpousal Support, 14 Intl Rev L & Econ 493 (1994); Margaret F. Brinig and Steven M.Crafton, Marriage and Opportunism, 23 J Legal Stud 869 (1994); Margaret F. Brinig andMichael V. Alexeev, Trading at Divorce: Preferences, Legal Rules and Transactions Costs,8 Ohio St J Dispute Resol 279 (1993); Martin Zelder, Inefficient Dissolutions as a Con-sequence of Public Goods: The Case of No-Fault Divorce, 22 J Legal Stud 503 (1993);Martin Zelder, The Economic Analysis of the Effect of No-Fault Divorce Law on theDivorce Rate, 16 Harv J L & Pub Pol 241 (1993); Allen M. Parkman, Unilateral Divorceand Labor Force Participation Rate of Married Women, Revisited, 82 Am Econ Rev 671(1992); Douglas W. Allen, Marriage and Divorce: A Comment, 82 Am Econ Rev 679(1992); Douglas Allen, An Inquiry into the State's Role in Marriage, 13 J Econ Behav &Org 171 (1990); H. Elizabeth Peters, Marriage and Divorce: Informational Constraints andPrivate Contracting, 76 Am Econ Rev 437 (1986); Yoram Weiss and Robert J. Willis,Children as Collective Goods in Divorce Settlements, 3 J Labor Econ 268 (1985); RobertA. Pollak, A Transaction Cost Approach to Families and Households, 23 J Econ Lit 581(1985); Yoram Ben-Porath, The F-Connection: Families, Friends, and Firms and theOrganization of Exchange, 6 Pop & Dev Rev 1 (1980); Elisabeth M. Landes, Economicsof Alimony, 7 J Legal Stud 35 (1978); Gary S. Becker, Elisabeth M. Landes, and RobertT. Michael, An Economic Analysis of Marital Instability, 85 J Pol Econ 1141 (1977);Theodore W. Schultz, ed, Economics of the Family: Marriage, Children and HumanCapital (Chicago, 1974).

6. See Allen M. Parkman, No-Fault Divorce: What Went Wrong? (Westview, 1992);Allen M. Parkman, Reform of the Divorce Provisions of the Marriage Contract, 8 BYUJ Pub L 91 (1993); Arthur B. Cornell, Jr., When Two Become One and Then ComeUndone: An Organizational Approach to Marriage and Its Implications for Divorce Law,26 Fam L Q 103 (1992); Michael J. Trebilcock and Rosemin Keshvani, The Role ofPrivate Ordering in Family Law: A Law and Economics Perspective, 41 U Toronto L J533 (1991); Ira Mack Ellman, The Theory of Alimony, 77 Cal L Rev 1 (1989); Gary S.Becker and Kevin M. Murphy, The Family and the State, 31 J L & Econ 1 (1988); LloydCohen, Marriage, Divorce and Quasi Rents; or, "I Gave Him the Best Years of My Life",16 J Legal Stud 267 (1987); Allen M. Parkman, The Recognition of Human Capital asProperty in Divorce Settlements, 40 Ark L Rev 439 (1987); Joan M. Krauskopf, Rec-

[2:517

Economics and the Problem of Divorce 519

approaches to family law, my conclusion is that the economic analysis ofdivorce presented in these recent works is not a sufficient basis for sweepingpolicy reform. The analysis is not useless or wrong, but it seriously underesti-mates the project at hand. Law may indeed promote efficiency in society, butlaw's work in the family is vastly more complex than normative economics hasbeen able to model. Indeed, this complexity is reflected in the divergentapproaches and recommendations in the current economic literature.

Section I of this Article describes the positive and normative economicunderstandings of marriage and divorce and explores two different ways inwhich divorce is described as an economic problem.

Section II demonstrates that the two different views of the problem ofdivorce lead to two different proposals for reforms. The first of the twoapproaches, that of Professor Becker and several other writers, conceives of theinterest in continuing a marriage as a type of property right and argues thatit should be transferable only with both parties' consent. I argue that thisformulation constructs divorce as a bargaining problem and applies Paretoefficiency norms with the goal of effecting optimal levels of divorce. Thesewriters recommend that divorce be available only with the mutual consent ofwife and husband. Problems within marriage are largely ignored, on theapparent assumption that these are less serious than the problem of divorce.

The second approach, more common in law and economics literature,emphasizes the goal of efficiency within marriage and argues for liability rulesfor divorce rooted in Kaldor-Hicks efficiency norms. Here, recommendationsfor divorce policy stress making the marital "contract" an enforceable one.This approach is less centered on the interest in remaining married and tendsrather to define the entitlements of marriage in financial terms.

The contrast between these solutions reveals different understandings ofefficiency and different concerns as to where family life should be rationalized.I argue that the merits of the two "solutions" therefore cannot be assessedwithout a broader normative frame of reference than this theory has devel-oped. My concern is greatest with those economic theories that treat theinterests of men and women in marriage and divorce as property rights.Although these theorists claim their solutions can improve the efficiency ofdivorce practices, they recognize only one of many important competingnormative interests in family and private life. I argue that, while problematic,proposals to create new liability rules to govern divorce may provide a betterframework for policy, because they are better able to accommodate a range ofdifferent interests.

In Section III, I argue that neither of these theoretical frameworks ade-quately accounts for most of the work of family law. Although economists

ompense for Financing Spouse's Education: Legal Protection for the Marital Investor inHuman Capital, 28 Kan L Rev 379 (1980); E. Raedene Combs, The Human Capital Con-cept as a Basis for Property Settlement at Divorce: Theory and Implementation, 2 JDivorce 329 (1979).

1995]

520 Roundtable

often argue that the central purpose of law in a market system is to compen-sate for the effects of market and contracting failures, this insight has not beenrecognized in the economic analysis of marriage and divorce. I argue thatprotecting individuals against the consequences of such failures in the familyis a central purpose of family law and suggest that whether these factors canbe adequately addressed in an economic analysis remains unresolved. In addi-tion, I argue that the economic models conceal a tension between "economic"and "moral" entitlements in family life. Any normative foundation for mar-riage and divorce law must confront this distinction. I believe this is a fun-damental problem with the efforts to reconstruct family law on the basis ofefficiency principles. Given these difficulties, I conclude that the project ofestablishing a Pareto optimal level of divorce or a set of efficient breach rulesfor marriage is not a useful goal for family law reform.

I. Economic Perspectives on Divorce

Economic theories of divorce rest on two pillars: a body of economicthinking about marriage, and a set of standard definitions of efficiency used ineconomic discourse. Generally, these theories depict divorce as a problemroughly analogous to the problem of efficient breach in contract theory.However, the efficiency problem in divorce is conceived in two distinctpatterns.

A. EcONOMICS AND FAMILY BEHAVIOR

Economists argue that their analytic tools are appropriate to the family aswell as the market because in both settings individuals make choices in a contextof scarce resources.7 In his Nobel lecture, Gary Becker explains the principlesthat have governed his investigations into family behavior:

The point of departure of my work on the family is the assumption thatwhen men and women decide to marry, or have children, or divorce, theyattempt to raise their welfare by comparing benefits and costs. So theymarry when they expect to be better off than if they remained single, andthey divorce if that is expected to increase their welfare.!

Initially, economic theorizing about the family remained close to the tradi-tional concerns of economics, exploring the connections between the householdand the larger economic system, and developing better analyses of human capital,labor force participation, and household production.9 More recently, economists

7. Sawhill, 106 Daedalus at 115-16 (cited in note 3); Blaug, Methodology of Econom-ics at 221 (cited in note 3).

8. Becker, 101 J Pol Econ at 395-96 (cited in note 1).9. I have explored some of these aspects of the "new home economics" in another

article. Ann Laquer Estin, Love and Obligation: Family Law and the Romance ofEconomics, 36 Wm & Mary L Rev 989 (1995).

[2:517

Economics and the Problem of Divorce 521

have become more ambitious, seeking to explain wider ranges of behavior withtheir models."0

This broader range of concerns has impelled a greater interest in law.Initially, this work was descriptive, focused on discovering latent economiceffects of (or explanations for) family law rules." However, writers in economicand legal publications increasingly have offered explicit policy arguments. In thiswork, economists model divorce decisions as problems of exchange and testoutcomes against efficiency norms. 2

Although there is a substantial literature in law and economics on efficientbreach of contract,13 the concept of efficient divorce is still startling. In legaltraditions, marriage and divorce regulation has been very different from the lawapplied to other varieties of contracts; 4 the "contract" idea has been used met-

10. See Blaug, Methodology of Economics at 221 (cited in note 3). This represents anincursion into terrain that was once the province of psychology or sociology and anthro-pology. For an assessment of this trend, see generally Ronald H. Coase, Economics and,Contiguous Disciplines, 7 J Legal Stud 201 (1978).

11. See, for example, Landes, 7 J Legal Studies 35 (cited in note 5).12. See, for example, Zelder, 22 J Legal Studies 503 (cited in note 5); Parkman, 8

BYU J Pub Pol L 91 (cited in note 6). There are two distinct models of family life in theeconomic literature. See Pollak, 23 J Econ Lit at 581-84 (cited in note 5). One is basedon Becker's household production theory, sometimes referred to as the "new home eco-nomics." This model depicts the household as a firm that combines the resources availableto it-the time of household members and various market goods-to produce the outputsor commodities it desires. See Becker, A Treatise on the Family (cited in note 3); Posner,Economic Analysis of Law at 139-41 (cited in note 2). A second model, rooted in"transaction cost" economics, is more directed toward exploring how marriage serves tostructure the complex, long-term relationships of husband and wife. The original develop-ment of transaction costs approach is credited to Oliver E. Williamson, Markets andHeirarcbies: Analysis and Antitrust Implications (Free Press, 1975). See also Oliver E.Williamson, Transaction-Cost Economics: The Governance of Contractual Relations, 22 JL & Econ 223 (1979).

Articles that apply a transaction costs approach to marriage and divorce includeCornell, 26 Family L Q 103 (cited in note 6), Pollak, 23 J Econ Lit 581 (cited in note5), and Ben-Porath, 6 Pop & Dev Rev 1 (cited in note 5). This model is implicit in otherwork, including writings describing marriage as a type of "relational contract." See Brinigand Crafton, 23 J Legal Stud 869 (cited in note 5); Cohen, 16 J Legal Stud at 271-73(cited in note 6), but see Ellman, 77 Cal L Rev at 28-32 (cited in note 6) (arguing thatthe relational contract analogy does not work because there are no common, externalstandards for marital relationships that courts could apply to guide resolution of disputes.)On relational contract theory, see generally Ian Macneil, The Many Futures of Contract,47 S Cal L Rev 691 (1974); Ian R. Macneil, Contracts: Adjustment of Long-Term Eco-nomic Relations under Classical, Neoclassical and Relational Contract Law, 72 Nw U LRev 854 (1978); Charles J. Goetz and Robert E. Scott, Principles of Relational Contracts,67 Va L Rev 1089 (1981). Under both approaches, however, marriage is understood asa type of contract, and the policy recommendations for divorce law place efficiency as acentral norm.

13. See notes 31-36.14. The legal literature is full of reinterpretations of the idea of "contract" in the

family setting. According to the United States Supreme Court, in 1888:[W]hilst marriage is often termed by text writers and in decisions of courts a civil

1995]

522 Roundtable

aphorically in family law, empty of most of the content it carries in othersettings."5 With the advent of no-fault divorce, however, marriage laws haveshifted toward market principles. 6

Some recent legal academic writing has urged greater freedom for parties toa marriage to set the terms, conditions, and duration of their marriages. 17 Thenew economic approaches to the family, with their foundation in marketprinciples, also suggest a much broader role for contract-based approaches tomarriage and divorce."8 These theories treat marriage not as a fixed category

contract . . . it is something more than a mere contract. The consent of the partiesis of course essential to its existence, but when the contract to marry is executedby marriage, a relation between the parties is created which they cannot change.

Maynard v Hill, 125 US 190, 210-11 (1888) (upholding constitutionality of an 1852legislative Oregon divorce in the context of a dispute over the property rights of ahusband's first and second wives).

Similar distinctions mark more contemporary case law as well; see, for example, In re theMarriage of Franks, 189 Colo 499, 506-07, 542 P2d 845, 850 (1975) (en banc) (rejectingargument that the no-fault divorce grounds in the Uniform Marriage and Divorce Actviolate the contracts clause of the state constitution); DeBurgh v DeBurgh, 39 Cal 2d 858,863-65, 250 P2d 598, 600-01 (1952) (rejecting analogy between recrimination defense todivorce and doctrine of mutual and dependent covenants in contract).

15. According to Judge Posner, three features set "traditional marriage law" apart fromcontract law:

First, the parties are not free to set the term of the contract or even toterminate the contract by mutual consent; the term is life, subject to termination forcause-much as in a university tenure contract.

Second, despite the long term of the contract, the sanction for breach is moresevere than in the case of a regular contract. If the husband abandons the wife (orvice versa), he not only must continue to support her (which is analogous to hishaving to pay damages) but may not marry anyone else. It is as if a contractbreaker could be enjoined from making a substitute contract for the one he hadbroken-and for the rest of his life.

Third, despite the locked-in nature of their relationship, if spouses have adispute during the course of the marriage the courts generally will not intervene tosettle the dispute; the spouses will have to work it out for themselves.

Marriage law is thus a puzzling amalgam of intrusiveness (in regard to theterm of the contract and the sanctions for breach) and hands-off-ness.

Posner, Economic Analysis of Law at 143 (cited in note 2) (footnotes omitted). See alsoCohen, 16 J Legal Studies at 271-73 (cited in note 6) (discussing marriage as contract).

16. Many writers have described a trend in family law toward greater use of contractand private ordering principles. See, for example, Jana B. Singer, The Privatization ofFamily Law, 1992 Wis L Rev 1443; Milton C. Regan, Jr., Market Discourse and MoralNeutrality in Divorce Law, 1994 Utah L Rev 605. Posner ties the shift toward moreliberal divorce laws to declining economic gains from marriage and the increasing costs ofenforcing strict divorce rules. Posner, Economic Analysis of Law at 141, 143-46. Thoughmarriage regulation is more market-based than was historically the case, older-status based-norms still have force. See Estin, 36 Wm & Mary L Rev at 1038-42 (cited in note 9).

17. Those urging a greater role for contract in marriage and divorce law includeLenore J. Weitzman, The Marriage Contract: Spouses, Lovers, and the Law (Free Press,1981), Jeffrey Evans Stake, Mandatory Planning for Divorce, 45 Vand L Rev 397 (1992);Gregg Temple, Freedom of Contract and Intimate Relationships, 8 Harv J L & Pub Pol121 (1985); Marjorie Maguire Shultz, Contractual Ordering of Marriage: A New Modelfor State Policy, 70 Cal L Rev 204 (1982).

18. See, for example, Peters, 76 Am Econ Rev at 437-39 (cited in note 5) (utilizing a

[2:517

Economics and the Problem of Divorce 523

with clearly defined legal and economic consequences, but as a relationshipsubject to continual renegotiation by its participants. At any point, breach is oneoption available to the parties, and the problem of divorce is fundamentally aproblem of remedies for breach of the marital contract.

The family law reforms that economists propose cast family obligations anddisputes in the mold of business contracts. These writers discuss the importanceof enforceable contractual rights in enabling a business (or marriage) to operateefficiently. 9 They view contractual protections as an incentive toward entry intoformal marriage and decisions to commit resources to the marriage."

Contractual ordering of marriage and divorce is appealing because of thebelief that it best accommodates a wide variety of preferences. The advocates ofcontract assert that it facilitates private choices about the family, from the eco-nomic structure of relationships to a broad range of lifestyle issues." In addi-tion, they argue that contract accommodates the shift away from permanence, byoffering a better mechanism for dispute resolution and divorce.'

From this perspective, divorce is not in itself a problem, just as breaches ofbusiness contract are not necessarily bad.' Rather, economists argue that thegoal of contract law and divorce law alike should be to encourage efficient

"contract-theoretic framework" to model marriage).19. See, for example, Cohen, 16 J Legal Studies at 296 (cited in note 6) ("[W]henever

contracts are not fully enforceable, an inefficient allocation of resources will follow.").20. Id. But see Ellman, 77 Cal L Rev at 23-24 (cited in note 6) (rejecting contractual

theory of marriage as inconsistent with no-fault divorce policies).21. See, for example, Shultz, 70 Cal L Rev at 247-48, 269-72, 329-30 (cited in note

17); Temple, 8 Harv J L & Pub Pol at 127 (cited in note 17). Carol Weisbrod drawsparallels to the legal realist and Uniform Commercial Code understanding of commercialcontracts to suggest that we might attempt to determine the "bargain in fact" underlyingvarious domestic arrangements. Carol Weisbrod, The Way We Live Now: A Discussion ofContracts and Domestic Arrangements, 1994 Utah L Rev 777, 807-13.

22. See, for example, Shultz, 70 Cal L Rev at 307-28; Temple, 8 Harv J L & Pub Polai 161-65. See also Trebilcock and Keshvani, 41 U Toronto L J at 549-50 (cited in note6). Writers disagree on what the effects of a shift toward contract would be. ElizabethScott argues that a more systematic use of contractual principles could help to foster"binding commitment and fulfillment of responsibility in family relations." Elizabeth S.Scott, Rehabilitating Liberalism in Modern Divorce Law, 1994 Utah L Rev 687, 689. CarlSchneider argues that the net effects of the move toward contract in family regulation hasbeen to release people from obligations historically embodied in law and social norms.Carl E. Schneider, Marriage, Morals, and the Law: No-Fault Divorce and Moral Discourse,1994 Utah L Rev 503, 532-35.

23. If divorce itself were the problem, economists would no doubt advocate abolishingit entirely. Although this position has some economic justification, it is generally rejectedbecause it would prevent "efficient" dissolutions of marriage as well as inefficient ones.Cohen, 16 J Legal Studies at 302 (The virtues include "extreme incentives for greater carein choosing a spouse."); Posner, Economic Analysis of Law at 144 (cited in note 2)("[M]aking divorce more difficult may actually foster happy marriages.").

199s]

524 Roundtable

allocations of resources.24 The obvious difficulty in both settings lies in identi-fying and regulating those breaches that lead to inefficient behaviors, whetherduring marriage or on divorce. In the process, the first step is to define efficiency.

B. EFFICIENCY THEORIES

Economists define efficiency in two ways. An efficient allocation of resourcesmay be defined as one in which their value is maximized, understanding that"value" is measured by how much an individual is willing to pay. This isdescribed as a Kaldor-Hicks approach." This method values the aggregate ofvalues in society rather than their distribution. Alternatively, the Pareto modeldefines an allocation as efficient when it results from the exchange of resourcesin a voluntary, informed transaction.26 Here, the consent of the parties indicatesthat both believe the transaction will make them either as well or better off thanpreviously.

Pareto efficiency norms follow two premises: that individuals are always thebest judges of their own utility, and that interpersonal comparisons of utility areimpossible in the absence of exchange.27 Decisions that are not the product ofvoluntary agreement may generate winners and losers, and because it is difficultto measure the impacts of those decisions on the affected individuals, it is hardto determine if there has been a net improvement in social welfare.2" The nor-mative preferences of economics dictate that as far as possible, relationshipsshould be based on exchange and consent.29 As Judge Posner argues, the condi-tions for determining Pareto efficiency must be stringent: an exchange is not avalid indicator of consent unless it is completely voluntary and has no effects onanyone external to the exchange.3"

In commercial settings, these norms dictate that exchange relationships

24. According to Lloyd Cohen: "Just as it is often efficient for contracts to bedissolved, so too both parties to a marriage may gain as a result of divorce." Cohen, 16J Legal Stud at 302 (cited in note 6).

25. Posner, Economic Analysis of Law at 13 (cited in note 2). Trebilcock andKeshvani posit that "Kaldor-Hicks efficiency ... would ask the question: would thiscollective decision (a change in legal rules, for example) generate sufficient gains to thebeneficiaries of the change so that they could, hypothetically, compensate the loserssufficiently to render the latter fully indifferent to it but still have gains left over for them-selves . . . ?" Trebilcock and Keshvani, 41 U Toronto L J at 534 (cited in note 6)(emphasis in original).

26. Posner, Economic Analysis of Law at 13-14. According to Trebilcock andKeshvani, "Pareto efficiency would ask of any transaction or policy or legal change: willthis transaction or change make somebody better off while making no one worse off?"Trebilcock and Keshvani, 41 U Toronto L J at 534.

27. Trebilcock and Keshvani, 41 U Toronto L J at 534.28. Id at 535.29. Posner suggests that consent to an exchange can be presumed where the exchange

is economically efficient. Richard A. Posner, The Economics of Justice 92-99 (Harvard,1981).

30. Posner, Economic Analysis of Law at 13-14.

[2:517

Economics and the Problem of Divorce 525

should be freely entered into and freely terminable, particularly where there areonly two parties involved; however, there must be either consent to terminate therelationship or an appropriate remedy for breach."' Contract law rules allowbreaches that are efficient under either criterion of efficiency: parties to acontract can agree to modify or rescind their contract, or, failing agreement, acourt can require that the party in breach compensate the other's losses fromnonperformance. In order for a breach to be Pareto efficient, both parties mustconsent to termination. In an action for damages, a judicially-determinedcompensation is substituted for consent to the contract termination. Economictheory suggests that remedies can promote efficient allocation of resources ifcompensation is determined in a manner reflecting Kaldor-Hicks efficiencynorms.

In normative terms, it is an ideal solution when parties to an agreementdetermine on what terms to go their separate ways, because by hypothesis theyare best able to judge their own utility. In commercial law, of course, contractbreaches are not always resolved by agreements to terminate. Economists defendthe role of court determinations of damages for breach of contract when comingto agreement is not feasible without extensive and costly negotiation. 2 Ineconomic terms, these damages discourage a contract breach or termination un-less the benefits to the breaching party are greater than the loss that breachimposes on the other.33 Thus, legal rules making contracts enforceable andproviding remedies for breach appear to be designed to further efficiency goalsrather than as a moral obligation to keep promises.3 4

These two sets of norms are not sharply delineated in contract law: the factthat contracts are legally enforceable through damages serves in part to bring theparties to a bargain over termination. In unusual cases, a court may orderspecific relief rather than damages, which has the effect of returning the matterto the parties for decision. In law and economics theory, the boundary betweendamages and specific relief in contract law is contested." The distinction reflectsa difference between, on the one hand, efforts to limit breach and protectreliance in individual cases, and on the other, efforts to permit breach and allowexchanges that will move resources to the place where they are the most usethroughout the economy. Thus, the economic literature on contract breach rec-ognizes that the choice of remedy may vary depending on whether excessive

31. Id at 117-20. But see Weisbrod, 1994 Utah L Rev at 778-79 (cited in note 21)(discussing the "perform or pay damages" approach to contract breach that is attributedto Holmes).

32. Judge Posner considers Kaldor-Hicks efficiency norms to be preferable here, becausehe views the transaction costs of applying Pareto standards to be too great. Posner,Economic Analysis of Law at 62, 119, 131 (cited in note 2).

33. The breach cannot be simply opportunistic, or an attempt to take advantage of thevulnerability of the nonbreaching party with no countervailing economic gain. Id at 117-18.

34. Id at 117-19. See also text accompanying notes 445-57.35. See, for example, Ian R. Macneil, Efficient Breach of Contract: Circles in the Sky,

68 Va L Rev 947 (1982).

1995]

526 Roundtable

breach or excessive performance seems to be a greater problem."6 The choicebetween normative efficiency principles reflects this difference in concern.

In the economic literature on divorce, there is a similar difference between aconcern with the efficiency of divorce decisions, which parallels the concern incommercial settings with excessive breach, and a concern with efficiency in socialinstitutions of marriage, which parallels the commercial concern with excessiveperformance. When these different understandings of efficiency are applied, theproblem of divorce is conceptualized in two distinct ways.

C. THE PROBLEM OF DIVORCE

Both normative economic approaches to divorce are rooted in the positiveeconomic theory of marriage pioneered by Gary Becker. In his Treatise on theFamily, Professor Becker discusses divorce in two significant ways: first, he ana-lyzes the circumstances in which an individual may gain from divorce; 7 andsecond, he provides a demographic discussion of the economic basis and effectsof changing divorce rates in modern society.38

Becker describes divorce decisions as an occasion for rationaldecisionmaking. In particular cases, decisions to divorce do not evidence failure;rather, "[d]issolution is a response ... to new information, favorable as well asunfavorable."39 Becker distinguishes between early and later divorces, arguingthat early divorces occur because potential marriage partners have only imperfectinformation about each other prior to marriage,4" and that later divorces are a"response to a demand for variety in mates or to life-cycle changes in traits."4

As Lloyd Cohen writes: "[o]ver time, the utility functions, information andopportunities of both marriage partners change." 42

Becker and others explain rising divorce rates in economic terms. Aswomen's earning power has increased, the value to a family of a wife's time hasalso increased, raising the "cost" of her nonmarket work. As raising children hasbecome relatively more expensive, there has been a dramatic decline in fertility.Taken together, the drop in birthrates and the increase in women's earnings havereduced the economic gain from specialization of labor in marriage.43

36. William Bishop, The Choice of Remedy for Breach of Contract, 14 J Legal Stud299 (1985); Robert Cooter and Melvin Aron Eisenberg, Damages for Breach of Contract,73 Cal L Rev 1432 (1985). June Carbone draws an analogy from this work to divorceremedies. See June Carbone, Economics, Feminism, and the Reinvention of Alimony: AReply to Ira Ellman, 43 Vand L Rev 1463, 1485-88 (1990).

37. Becker, A Treatise on the Family at 331-41 (cited in note 3).38. Id at 350-361.39. Becker, Landes, and Michael, 85 J Pol Econ at 1151-52 (cited in note 5). The

authors continue: "Moreover, the freedom to dissolve reduces the impact of unfavorableinformation, and thereby reduces the incentive to delay marriage or otherwise search morein order to avoid a mismatch." Id at 1152.

40. Becker, A Treatise on the Family at 327-28.41. Id at 327. See also Becker, Landes, and Michael, 85 J Pol Econ at 1143.42. Cohen, 16 J Legal Stud at 267 (cited in note 6).43. Becker, A Treatise on the Family at 350-55 (cited in note 3); Posner, Economic

[2:S17

Economics and the Problem of Divorce 527

This explanation of divorce suggests that marriage is caught in a viciouscircle. The increasing likelihood of divorce has consequences for family behavior,such as reducing fertility, "because children are more difficult to rear and mayprovide less pleasure after a marriage dissolves."' Moreover, these factors leadwomen to increase their labor force participation and investment in market-oriented human capital as a hedge against divorce.45 This in turn further reduc-es the economic gain from marriage.

These positive analyses take an agnostic stance, offering no opinions onwhether the labor market changes that have driven or accompanied changes inthe social institution of marriage are good or bad, or whether individual deci-sions to divorce are right or wrong. In more recent normative work, however,Becker and other economists have recommended that laws be changed to makedivorce more difficult. 6 Their argument for limits on divorce is formulated inapparently neutral terms: divorce should be freely permitted as long as it isefficient. Their criticism of current no-fault divorce laws is that many divorcesare not efficient; they recommend that the law be designed to prevent suchdivorces. They define an "efficient" divorce as one in which the combined benefitto the husband and wife from divorce is greater than their combined benefitfrom remaining married.4"

Another recent argument posits that divorce should only be permitted onterms which encourage efficient arrangements within marriage. This is a premiseof work by Lloyd Cohen, Ira Ellman, and others who argue that these factorsshould be the basis of financial remedies made available to spouses at the timeof divorce.4" An "efficient" household is one in which there are gains from aspecialization and division of labor.49 The rationale of this work is that changesin the legal entitlements that spouses acquire as a result of marriage could helpstem or reverse the general economic trends away from specialization of labor inmarriage.50

Analysis of Law at 141 (cited in note 2) ("The declining marriage rate and rising divorcerate suggest that the benefits of marriage have been falling relative to the costs."); Becker,Landes, and Michael, 85 J Pol Econ at 1180-81 (cited in note 5). For a detailed demo-graphic analysis, see Andrew J. Cherlin, Marriage, Divorce, Remarriage: Revised andEnlarged Edition (Harvard, 1992).

44. Becker, A Treatise on the Family at 355.45. Id at 355-56. See generally Kathleen Gerson, Hard Choices: How Women Decide

about Work, Career and Motherhood (California, 1985) (detailing factors that influencewomen's decisions concerning domestic or market orientation).

46. See, for example, Parkman, No-Fault Divorce at 137-40 (cited in note 6); Becker,Bus Wk at 22 (Dec 7, 1992) (cited in note 4); Zelder, 16 Harv J L & Pub Pol at 242(cited in note 5). I argue below that the normative economic policy argument for newlegal restrictions on divorce has not been well synthesized with the positive economic ex-

planation of increasing divorce rates; see text accompanying notes 180-90.47. See, for example, Landes, 7 J Legal Stud at 36-39 (cited in note 5).48. See, for example, Cohen, 16 J Legal Stud at 302-03 (cited in note 6); Ellman, 77

Cal L Rev at 50-51 (cited in note 6).49. Becker, A Treatise on the Family at 30-48.50. See, for example, Becker, Bus Wk at 22 (Dec 7, 1992) (cited in note 4).

19951

528 Roundtable

Both models challenge the policies which have made marriage effectivelyterminable at will by either partner. As many writers have noted, these rules haveeffectively converted the traditional marriage contract into something more akinto a traditional at-will employment contract"' or to a voluntary contractualassociation terminable at will by its members. s2 Yet while economists are gener-ally supportive of employment at will, 3 marriage at will appears to be a signifi-cant social problem. One concern is that if the divorce decision can be made byone partner, acting alone, that partner will not typically take into account thefull range of costs and benefits of marriage and divorce for both parties. Thus,divorce may occur even when the parties' combined gain from marriage is greaterthan their combined gain froin divorce. 4 Divorce in these circumstances alsowill generate distributive problems: one partner can appropriate a greater shareof the parties' combined gain by choosing divorce, particularly where there areno effective financial remedies in place."s A third concern is that no-fault di-vorce discourages economically efficient marriages, set up along traditionalgender role-divided lines, by failing to compensate women for their contributions

51. For a description of contemporary divorce laws as creating marriage at will, seeStephen D. Sugarman, Dividing Financial Interests on Divorce, in Stephen D. Sugarmanand Herma Hill Kay, eds, Divorce Reform at the Crossroads 130, 138-39 (Yale, 1990);Lenore J. Weitzman, The Divorce Revolution 376 (Free Press, 1985); Mary Ann Glendon,The New Family and the New Property 199-201 (Butterworth, 1981).Allen Parkman criticizes this point, arguing that an analogy to a "commercial contract,"

with a requirement of compensation in the event of a breach, is a more "appropriate"analogy. Parkman, No-Fault Divorce at 128, 147 n 50 (cited in note 6). However, thisseems to miss the point, which is that the at-will analogy more accurately captures thecurrent law. Sugarman, Dividing Financial Interests at 138-39.Where there is at-will employment, one party has no legal remedy if the other chooses

to terminate the relationship. A remedy would only be available for breach of theagreement before termination. See, for example, Margaret F. Brinig, The Law andEconomics of No-Fault Divorce: A Review of No-Fault Divorce: What Went Wrong? byAllen M. Parkman, 26 Fain L Q 453, 468 (1993) (arguing that "if the parties breach acontract at will before termination, there is compensation for the breach").

52. The analogy appears in Posner, Economic Analysis of Law at 142-43 (cited in note2); Cornell, 26 Fam L Q at 116-20 (cited in note 6); Ellman, 77 Cal L Rev at 13-14(cited in note 6), and, at a more metaphoric level, in dozens of case decisions and law re-view articles. For more extended discussion, see Cynthia Starnes, Divorce and theDisplaced Homemaker: A Discourse on Playing with Dolls, Partnership Buyouts andDissociation under No-Fault, 60 U Chi L Rev 67 (1993); Bea Ann Smith, The PartnershipTheory of Marriage: A Borrowed Solution Fails, 68 Tex L Rev 689 (1990).

53. See Richard A. Epstein, In Defense of the Contract at Will, 51 U Chi L Rev 947(1984) (arguing that over a very broad range of circumstances, at-will employment worksto the benefit of both employer and employee at the time their contract is formed); seealso Cohen, 16 J Legal Stud at 299 (cited in note 6) (noting that at-will contracts are notused in commercial contexts in which parties make differential investments, because therisks to one party from opportunistic behavior-breach-are too great). It is ironic thatthe shift in divorce law has taken place as employment law has begun to shift towardincreasing the protections afforded to workers. See Glendon, The New Family at 151-170.

54. See text accompanying notes 85-87.55. See text accompanying notes 151-58 and 266-98.

[2:517

Economics and the Problem of Divorce 529

in the home.5'While sharing these concerns, economists and lawyers influenced by econom-

ics reach different policy recommendations. Some scholars, including ProfessorBecke; call for abolishing unilateral divorce, so that marital freedom could beachieved only by an exchange of consents. In effect, this group argues thatdivorce should be permitted only when the stricter Pareto criterion of efficiencyis met. Others would allow divorce more freely, but in the context of new rulesthat create incentives and protections for certain "efficient" behaviors in mar-riage. These proposals would implement something closer to the Kaldor-Hicksefficiency norm, by making marriage contracts enforceable through determina-tions of breach and damages. Both approaches would protect certain entitlementsin marriage, but the former would do so with a property rule, the latter with aliability rule. 7

D. PROPERTY RULES AND LIABILrrY RULES

Law and economics theory begins with some ideas about the relationshipbetween law and economics. At one level, theorists explore the role of law infacilitating (or imitating) market relationships.5" At another level, economictheorists seek to explain legal rules through economic reasoning. 9 At a deeperlevel, theorists recognize that law is the background and foundation for econom-ic activity; our system of voluntary exchange commonly modeled in economictheory is built on a foundation of property, contract, and tort law."0 In theirexplorations of legal rules, economic theorists distinguish two important ques-tions: what entitlements the law will grant or recognize, and how thoseentitlements will be legally protected."1 Guido Calabresi and A. DouglasMelamed contrast three types of rules used to protect various entitlementsthroughout the common law system: property rules, liability rules, and inalien-ability rules."2 Their elaboration of these concepts makes clear that the three

56. See text accompanying notes 120-24 and 205-16.57. Eric Talley has pointed out to me that the usual understanding of these concepts

is that Pareto efficiency "nests" the Kaldor-Hicks variety. See text accompanying notes 69and 72-80.

58. Posner, Economic Analysis of Law at 21 (cited in note 2).59. Id at 23-25.60. See Michael Trebilcock, The Limits of Freedom of Contract 9-10, 15-17 (Harvard,

1993); Posner, Economic Analysis of Law at 31.61. Guido Calabresi and A. Douglas Melamed, Property Rules, Liability Rules, and In-

alienability: One View of the Cathedral, 85 Harv L Rev 1089 (1972).62. They distinguish three means of protecting entitlements:

An entitlement is protected by a property rule to the extent that someonewho wishes to remove [it] from its holder must buy it from him in a voluntarytransaction in which the value of the entitlement is agreed upon by the seller....

Whenever someone may destroy the initial entitlement if he is willing to payan objectively determined value for it, an entitlement is protected by a liabilityrule....

An entitlement is inalienable to the extent that its transfer is not permittedbetween a willing buyer and a willing seller.

1995]

530 Roundtable

methods have significantly different effects. Entitlements protected by propertyrules can be shifted only in a consensual transaction-the domain of contract.Although property rules define entitlements, the rules do not attempt to specifythe value of these rights, which is normally left to the subjective measurement oftheir holder.63 Entitlements protected by liability rules can be shifted withoutthe consent of their holder, subject to compensation-the usual domain of tort.Liability rules contemplate that the value of these entitlements may need to bedetermined on some objective ground by some collective process.64 Inalienabilityrules define entitlements that cannot be shifted."' Calabresi and Melamedreview a series of considerations that come into play in settingentitlements-some derived from economic efficiency criteria, and some not.Ideally, the initial choice of an entitlement would lead to optimal allocations ofresources, taking into account various types of transaction costs and how readilythey can be avoided under different rules. It would also take into account the ef-fects of different entitlements on distributions of wealth, as well as external orthird party effects and existence of "moralisms."66

For Calabresi and Melamed, the choice between property and liability rules,and between two entitlements, requires a careful assessment of these costs. Theysuggest that property rules will be preferable whenever an entitlement may bereadily exchanged in a privately negotiated transaction. But in "situations wheremarkets may be too expensive or fail," Calabresi and Melamed argue thatcollective valuation and transfer through liability rules may be more efficient.67

They also point out that liability rules may be employed "to favor distributivegoals" that would be difficult to achieve under property rules.6"

This distinction between property rules and liability rules in many respectsparallels the distinction between Pareto and Kaldor-Hicks efficiency norms.69

Pareto norms imagine a voluntary transfer, an exchange of entitlements betweentwo participants who have made independent determinations of the values atstake. Kaldor-Hicks norms, on the other hand, imply the sort of collectivevaluation process that is characteristic of liability rules.

The questions of what entitlements the law should grant and how those

Id at 1092.63. Calabresi and Melamed note that "most entitlements to most goods are mixed."

Id at 1093. For example, an entitlement to real estate "may be protected by a propertyrule in situations where [a buyer] wishes to purchase it, by a liability rule where thegovernment decides to take it by eminent domain, and by a rule of inalienability insituations where [the owner] is drunk or incompetent." Id.

64. Id at 1092.65. Id at 1092-93.66. Id at 1093-1105. For discussion of transaction costs and third party effects, see

text accompanying notes 423-39.67. Id at 1107-10 (cited in note 61). See also Posner, Economic Analysis of Law at

33-35 (cited in note 2); Posner, Economics of Justice at 69-72 (cited in note 29) (regard-ing exclusivity of property rights).

68. Calabresi and Melamed, 85 Harv L Rev at 1110.69. See text accompanying notes 25-34.

[2:517

Economics and the Problem of Divorce 531

entitlements should be protected are extremely important and deserve a centralplace in these debates. Calabresi and Melamed provide a framework whichinvites the contributions of law and economics into both inquiries, as well as intothe larger methodological questions suggested by the different efficiency norms.So far, however, the invitation has not been accepted. Instead, the economicanalysis of divorce has split into two approaches: the argument for mutualconsent rules, which is an argument based on a property interest in maritalstatus,70 and the argument for contract enforcement remedies, which is an argu-ment based on liability rules. 7' Although the difference between property rulesand liability rules has enormous practical and theoretical significance, it has gonevirtually unacknowledged.

II. Two Solutions to the Problem of Divorce

Two different understandings of the problem of divorce correspond with twodifferent, and inconsistent, solutions. Both solutions are based on argumentsfrom economic efficiency, but they conceive and apply these norms differently.One would allow divorce only on mutual consent grounds, leaving the specifica-tion of rights and responsibilities entirely to the parties. The other would providefor unilateral divorce based on a judicial determination of breach, and accompa-nied by orders for payment of damages.

"Mutual consent" proposals, rooted in a Pareto efficiency model, reflect theconcern that no one should be made worse off individually as a result of di-vorce.7 The proposals build upon a variety of empirical and theoretical studiesof divorce decisions under laws which do not require mutual consent. 73 By de-

70. This is explicit in the literature; see text accompanying note 113. In his discussionof contract remedies, Macneil describes specific performance as a property rule and expec-tation damages as a liability rule. Macneil, 68 Va L Rev at 962-63 (cited in note 35).

71. Some discussions attempting to apply the Coase Theorem to divorce rules contrasta property interest in marriage with a property interest in divorce. See text accompanyingnote 113. My question is whether there is a sound basis for either form of property rule.See text accompanying notes 114-18 and 312-48.

72. A divorce freely agreed to by both husband and wife would be Pareto superior.By definition, the parties' consent means that neither is being made worse off. The premiseof mutual consent rules is that a party who stands to suffer from divorce will bargainover the price of consent. If the gain from divorce is great enough, the partner will beable to pay the price of consent. On the other hand, if the price of consent is more thanthe spouse can afford, the divorce will not occur. Becker acknowledges that a divorcemight be blocked despite generous terms by a "bitter or spiteful spouse" who refuses con-sent; this is one reason he advocates allowing arbitrated divorces after several years'separation. Becker, Bus Wk at 22 (Dec 7, 1992) (cited in note 4).

73. Economists dispute whether changes in the legal rules of divorce substantially affectthe rates of divorce. See Parkman, No-Fault Divorce at 75-79 (cited in note 6). As anapplication of the Coase Theorem, mutual consent and unilateral divorce rules would beequivalent if the parties had equal and unbiased information and if all aspects of martialincome could be costlessly transferred. Landes, 7 J Legal Stud at 38-39 (cited in note 5).These conditions are not met in the setting of divorce, however. Id at 39.40. Even if theCoase Theorem applied, however, shifts in the rules would have distributional consequenc-

19951

532 Roundtable

nying divorce where there is no agreement, this proposal is the equivalent of anargument for enforcement of marital obligations through specific performanceremedies. These proposals define the entitlements of marriage as property rights,and thus accord each partner ownership of the other's marital status. In otherwords, the parties must stay married until they agree otherwise.

"Contract enforcement" proposals, similar to the efficient breach theoriesused in business contract settings, follow a Kaldor-Hicks standard of efficiency.Enforcement proposals purport to foster efficient decisions during marriage.Although these theories also assert that divorce should be discouraged unlessthere is sufficient gain to the breaching party to offset the other's losses, themechanisms relied upon to encourage efficient marital decisions and discourageinefficient divorces are liability rules which provide financial recoveries to theparty who would prefer to remain married.74 As with contract damages, theseare intended to approximate that party's losses from divorce, particularlyeconomic losses.

These policy recommendations are difficult to compare, because they expressdifferent decisions as to what entitlements society should recognize and howthose entitlements should be protected.7" In effect, mutual consent proposalsplace primary normative concern on divorce, while contract enforcement propos-als place this concern on marriage.

In theoretical work, the divide between these approaches is often minimized.The concept of Pareto efficiency is understood to "nest" the Kaldor-Hicksvariety, so that results that satisfy the Kaldor-Hicks test are also understood aspotentially Pareto optimal.7 Liability rules are described as efforts to replicatethe results that an actual bargaining process would yield if it were feasible.77

Moreover, the Coase Theorem suggests that the legal rules governing divorcemay not affect the levels of divorce in society, at least where all aspects ofmarital income are completely divisible and can be costlessly transferred betweenhusband and wife.78 Even then, however, different divorce rules would result in

es. See Peters, 76 Am Econ Rev at 438 (cited in note 5); Allen, 82 Am Econ Rev at 683-84 (cited in note 5); Zelder, 22 J Legal Stud at 504-06 (cited in note 5) (arguing thatbecause children are not transferable, no-fault divorce increases the divorce rate).

74. See, for example, Brinig and Crafton, 23 J Legal Stud at 892-94 (cited in note 5).See also text accompanying notes 193-98 and 205-42.

Lloyd Cohen has written to me that he disagrees with the distinction as I havedrawn it here. He argues that although "a property rule is of necessity Paretian" given theneed for the parties' consent, liability rules should be based on a Pareto standard as well.Because it is possible to compensate the losers in divorce, he writes, "efficiency and justicecall for full compensation" and "no economist should be talking Hicks-Kaldor as anormative standard." Letter from Lloyd Cohen to Ann Laquer Estin (Mar 23, 1995) (onfile with author). Nevertheless, Judge Posner relies primarily on Kaldor-Hicks norms in hisanalysis of contract remedies, see note 32, and Zelder describes his framework in theseterms as well, see note 112.

75. See text accompanying notes 312-48. This argument uses the framework developedby Calabresi and Melamed. See text accompanying note 62.

76. See text accompanying notes 25-26.77. See text accompanying notes 230-32.78. See, for example, Landes, 7 J Legal Stud at 36 (cited in note 5) (citing R. H.

[2:517

Economics and the Problem of Divorce 533

varying distributions of wealth between husband and wife after divorce.79 Also,these results will be influenced, perhaps dramatically, by externalities, transac-tions costs, and other types of market and contracting failures. 0 Moreover,analysis based on economic theory and large empirical samples is not well suitedto explore the practical effects that different legal frameworks may have forindividuals facing divorce decisions."'

A. MUTUAL CONSENT PROPOSALS

Many economic theories of divorce conceptualize the process as an exchangeor bargaining problem. Where there is sufficient gain from divorce, the gain canbe shared between the parties so that neither is made worse off as a result of thedivorce." Mutual consent proposals envision that this allocation results fromnegotiation and agreement between the two divorcing parties. Absent agreement,there is no ground for divorce.8 3 This is Professor Becker's position: "[T]he no-fault experiment should be abandoned and replaced by laws that allow divorceonly when both the husband and wife agree-what is called divorce by mutualconsent."84 Changing divorce laws to limit dissolution to circumstances of

Coase, The Problem of Social Cost, 3 J L & Econ 1 (1960)); see also Allen, 82 Am EconRev at 683-84 (cited in note 5); Peters, 76 Am Econ Rev at 437-38 (cited in note 5). Inaddition, strategic behavior and various contracting and market failures can distort thisoutcome. See generally A. Mitchell Polinsky, An Introduction to Law and Economics 11-14 (Little, Brown, 2d ed 1989); Zelder, 22 J Legal Stud at 504 n 3 (cited in note 5).

79. Peters, 76 Am Econ Rev at 438. In some arguments, including Becker's, thesedistributional consequences are the basis for the perception that inefficient divorce is asignificant social problem. See text accompanying notes 121-23.

80. A number of writers have elaborated factors that may lead to inefficient divorcedecisions under the present law, such as the presence of transactions costs and externalitiesand the impossibility of transferring public goods. See, for example, Zelder, 22 J LegalStud 503; Allen, 82 Am Econ Rev at 683-84. Both Allen and Zelder challenge Peters'sempirical finding that levels of divorce were not affected by the shift to no-fault rules.Allen argues that there are necessarily transaction costs whenever property rights at issuecannot be perfectly defined: "If property rights are not perfect, then resources are devotedtoward their protection and capture, and transaction costs are positive." Id at 683. Healso argues that Peters's finding that divorce results in systematic wealth transfers fromwives to husbands implies the same conclusion. Id at 683-84. Zelder's arguments are dis-cussed in the text accompanying notes 101-18.

81. In general, the practical legal aspects of these transactions are not well consideredin the literature, which is prone to broad generalizations about the details of divorceprocess. One notable exception is Brinig and Crafton, 23 J Legal Stud 869 (cited in note5) (discussing implications of different legal regimes for divorce in the context of spousalabuse). See generally text accompanying notes 351-55 and 409-17.

82. An illustration of how this negotiation might proceed is given in Ira Mark Ellman,Should The Theory of Alimony Include Nonfinancial Losses and Motivations?, 1991 BYUL Rev 259, 296-300.

83. Becker, Bus Wk 22 (Dec 7, 1992) (cited in note 4). Becker apparently wouldauthorize "binding private divorce arbitration" if there had been a failure to agree onterms after several years of separation. See note 72.

84. He made this claim in a column in Business Week magazine, published shortly

1995]

534 Roundtable

mutual consent offers two apparent advantages. First, these writers believe thatforcing bargaining between the parties will improve the outcomes of individualdivorce cases, moving them closer to an optimum result. Second, they stress thatit privatizes the process of divorce, eliminating the courts and, by implication,acrimony, expense, and delay.

1. Efficient divorce.

As noted above, the test of efficient divorce is whether the parties' combinedgain from marriage is less than their combined gain from divorce.8" The expect-ed gain or "wealth" referred to here is not conceived in narrowly financial terms;it includes all of the personal satisfactions and dissatisfactions of family life.Becker and his colleagues argue that if mutual consent were required, the partieswould agree to separate only if the party who expected to be less well off afterdivorce was compensated by the other.86 Moreover, because each partner ispresumptively best situated to determine his or her own preferences and utility,a mutual consent system seems ideal to take all of the intangible benefits ofmarriage or divorce into account. Allen Parkman argues that mutual consentwould lead to compensation for many factors that the current system discounts,including human capital issues, lost companionship, search costs of finding a newspouse, and disruption in the lives of children. 7

By this definition, an "inefficient divorce" is one that occurs even thoughhusband and wife together enjoy a greater gain from marriage than they wouldfrom divorce. Conversely, although this term is never used, an inefficient mar-riage must be one in which the husband and wife would have a greater combinedgain from divorce than they enjoy within marriage. Under a unilateral no-faultdivorce law, these inefficient marriages would be relatively rare; under mutualconsent rules, they could be quite common.

The argument for requiring mutual consent for divorce reflects a judgmentthat preventing inefficient divorces is more important than allowing inefficientmarriages to continue. But this conclusion must necessarily be derived from someargument beyond the scope of efficiency principles. Some accounts seek toeliminate any subsidiary normative judgment by basing the preference for oneregime on the conclusion that efficiency goals are more readily achieved in onecontext than the other.8

after he received his prize, without providing any technical microeconomic explanation. Id.85. Becker, Landes, and Michael, 85 J Pol Econ at 1144 (cited in note 5) ("[A] couple

would separate if, and only if, their combined wealth from remaining married wereexpected to be less than their combined wealth when separated.").

86. Id. Conversely, "if all compensations between spouses were feasible and costless,"a rule allowing unilateral divorce would have the same result, as the parties would agreeto stay married only if the party who expected to be better off in the marriage compen-sated his or her partner. Id.

87. Parkman, No-Fault Divorce at 138 (cited in note 6). For discussions of humancapital, see Becker, 101 J Pol Econ at 392-93 (cited in note 1); Estin, 36 Wm & Maryat 1006 (cited in note 9).

88. See, for example, the discussion of Zelder in text accompanying notes 101-18.

[2:517

Economics and the Problem of Divorce 535

Significantly, economic analyses of divorce have not distinguished betweenthe problems of achieving efficiency in marriage and efficiency in divorce. Mostoften, this failure follows from the assumption that these two do not conflict.That is, commentators have implied that a legal framework fostering efficientdivorce decisions will lead couples to make efficient marital decisions as well.There are, however, two distinct groupings among theorists who advocate anemphasis on efficient divorce. One concentrates on the exchange aspect ofdivorce, with the goal of fostering bargaining between husband and wife.Another group focuses on the outcomes of marriage and divorce bargaining, withthe express goal of remedying the social problems that result from presentdivorce practices.

a) Exchange models. Bargaining models of divorce are premised on the goalof reducing divorce levels to an optimum, defined as what would result ifnegotiations occurred under the ideal conditions, including perfect and symmetricinformation, no public goods, no transaction costs, and willing participants. Thisideal is far removed from the actual circumstances of divorce bargaining, but itserves as a theoretical base from which to explore the effects of the factors thatcomplicate divorce decisions.

Some studies investigate different transaction costs involved in divorce,ranging from long-term financial and psychological costs for husband andwife9 to social stigma and the time and money spent on litigation or bargain-ing." Other work models this question in a broader fashion by exploring coststhat may prevent parties from reaching optimal allocations of resources withinmarriage. 9'

Informational constraints may create another barrier to efficient divorces.Elizabeth Peters describes the problem: "Each spouse may not be able to deter-mine the value of opportunities at divorce that face the other spouse. Incentivesthen exist for each spouse to misrepresent the value of alternatives that he or shefaces and engage in costly strategic bargaining." 92

89. See, for example, Parkman, No-Fault Divorce at 122-23 (cited in note 6) (dis-cussing "costs of lost companionship, searching for an alternative spouse or companion,and disrupting the lives of children").

90. See, for example, id at 77-78; Brinig and Alexeev, 8 Ohio St J Dispute Res at287, 291 (cited in note 5); Zelder, 16 Harv J L & Pub Pol at 250 (cited in note 5).

91. Douglas Allen argues that the purpose of state intervention in marriage and divorceis to increase the gain to marriage by reducing a variety of transaction costs. Allen, 13 JEcon Beh & Org at 190-21 (cited in note 5). He argues that mutual consent and fault-based divorce rules reduce transaction costs by allocating "property rights over theduration of the marriage" to the spouse with fewer assets. Id at 184 (citing Gallaspy vGallaspy, 459 S2d 283 (Miss 1984)). In the absence of this protection, state laws haveresponded with new devices to mitigate these costs, such as equitable distribution rules andrules that permit consideration of fault in determining financial remedies. Allen, 13 J EconBeh & Org at 184.

92. Peters, 76 Am Econ Rev at 438, 442 (cited in note 5). Professor Peters's empiricaldata tend to support a model based on symmetric information, id at 450, leading her to

199s]

536 Roundtable

Some writers have discussed the external effects of divorce decisions, includ-ing in particular the effects of divorce on children." Children have been under-stood as another obstacle to efficient divorce: like other intangible benefits ofmarriage, children are "public goods" which cannot be readily exchanged or di-vided between husband and wife. As Elisabeth Landes writes, "if a large compo-nent of marital income for both spouses comes from the enjoyment of children,it would be impossible to transfer this income from one spouse to the other."94

This problem has also been explored by Yoram Weiss and Robert Willis," andby Martin Zelder 6

Arguments for exchange models often assume that present levels of divorceare too high, and formulate the central policy question as whether different legalstructures would cause divorce rates to go down. There is particular interest inwhether no-fault divorce rules are in some manner correlated with higher divorcerates observed in recent years. 7 Many authors have claimed that the fault-baseddivorce laws in existence prior to the 1970s operated as a mutual consentsystem, and that divorce rates from that era indicate what would occur under amutual consent system.9s Some writers conclude that a shift to mutual consentrules would reduce the incidence of divorce.

From the perspective of legal policy and history, the simple equation betweenfault-based and mutual consent divorce is a surprising one. Certainly the two

conclude that divorces do occur under no-fault rules when they are efficient, id at 453.93. See, for example, Posner, Economic Analysis of Law at 143-45 (cited in note 2);

Zelder, 16 Harv J L & Pub Pol at 259 n 32 (cited in note 5).94. Landes, 7 J Legal Stud at 39 (cited in note 5). She adds this insight:This may help to explain the commonly held notion that poor families tend toexperience a greater rate of desertion than middle and upper income families. Thesmaller the amount of money wealth held by a couple, the greater may be theproportion of total marital wealth that is derived from public goods such aschildren. Since transfers would be more difficult under these conditions, one wouldexpect to find a greater incidence of desertion among low income families, unrelatedto the legal costs of obtaining a divorce.

Id.95. Weiss and Willis, 3 Labor Econ 268 (cited in note 5).96. Zelder, 16 Harv J L & Pub Pol at 252. According to Zelder, other such public

goods include home heating and love. Id at 254 n 21. See also Zelder, 22 J Legal Studat 505-06 (cited in note 5).

97. See, for example, Zelder, 16 Harv J L & Pub Pol 24; Allen, 82 Am Econ Rev679 (cited in note 5); Peters, 76 Am Econ Rev 437.

98. See, for example, Zelder, 22 J Legal Stud at 504; Allen, 13 J Econ Beh & Orgat 182-83 (cited in note 5); Peters, 76 Am Econ Rev at 446. All of the empirical workin the literature appears to assume that data taken from states with fault-based rules fordivorce can be relied upon to indicate how mutual consent rules would affect divorcerates. The data also focus almost exclusively on grounds for divorce, without muchconsideration of the effects of specific alimony, property division, and custody rules. Butsee Brinig and Crafton, 23 J Legal Stud at 892-94 (cited in note 6); Peters, 76 Am EconRev at 441. For an empirical analysis suggesting that these rules are more significant thanthe change in grounds for divorce, see Marsha Garrison, The Economics of Divorce:Changing Rules, Changing Results, in Sugarman and Kay, eds, Divorce Reform at theCrossroads 75, 90-93 (cited in note 51).

[2:517

Economics and the Problem of Divorce 537

systems possess importantly distinct moral contents and generate very differentpractical consequences. While divorce agreements were regularly negotiatedunder fault-based divorce rules, this process involved substantial financialexpense and inflicted significant social and personal costs. These costs might wellhave reduced divorce rates to inefficient levels.99 None of the theoretical or em-pirical work has developed a method that allows us to test in an individual casewhether divorce would be efficient, or in the aggregate whether divorce isoccurring at an optimal level.' 0

With a pure exchange model, the only normative goal is to increase thescope of exchange. This is most explicit in an analysis by Martin Zelder, whoargues that divorce rules should be structured in order to expand the scope ofbargaining over the allocation of gains from marriage or divorce."'1 Zelderviews bargaining over divorce terms as preferable to bargaining over terms forcontinuation of a marriage, because some of the gain from marriage is in theform of public goods which cannot be traded between the parties. 2 He pres-ents empirical evidence that divorce rates have increased under no-fault rules andthat the reason for this increase is "the nontransferability of children withinmarriage."103 Zelder makes the normative judgment that these additionaldivorces are economically inefficient, that they would not occur in a system offault-based divorce.' Therefore, he says, no-fault divorce should be eliminated"so that marriage will be preserved when it is economically efficient.""05

Although Zelder's recommendation is couched in the neutral and mathemati-cal language of efficiency, his discussion implicitly depends on prior normativejudgments. Zelder argues that under a mutual consent or fault system, thepartner who wants a divorce must offer favorable terms in order to induceconsent. Under no-fault rules, however, the partner who wants to remainmarried must offer a greater benefit within marriage.' Zelder posits that thisis significantly harder to do, at least where a couple devotes much of theirfinancial resources to their children.0 7 That is, it is more difficult for onespouse to purchase the other's continued commitment to a marriage than topurchase consent to a divorce. Thus, the transaction is less likely to take place.Without a deal, divorce will result.

But why is this a problem? In Zelder's analysis, divorce on these facts is"inefficient"; that is, there are gains to one party from the marriage that would

99. These issues are noted in Posner, Economic Analysis of Law at 145-46 (cited innote 2). See also Homer H. Clark, Jr., The Law of Domestic Relations in the UnitedStates 409-11 (West, 1988).

100. See text accompanying notes 111-12.101. Zelder, 22 J Legal Stud at 504-05 (cited in note 5).102. Id at 507-09; Zelder, 16 Harv J L & Pub Pol at 252-54 (cited in note 5).103. Zelder, 16 Harv J L & Pub Pol at 257.104. Id at 258-59.105. Id at 262.106. Zelder, 22 J Legal Stud 504-05 ("[F]or example, a wife may agree to be less

critical of her husband in order to induce him to remain married.").107. Id at 505-06. See also note 94.

1995]

538 Roundtable

be lost in divorce. By definition, however, the problem only occurs when thesegains consist of shared goods, such as love or children, which are unequallyvalued by the partners to the marriage."0 8 Zelder's model ignores that requiringthe unhappy partner to buy out of a marriage in these circumstances is not asimple matter: it requires that the parties put a price on precisely those aspectsof marriage that are most difficult to value, most at risk from commodification,and to which they have attributed widely different valuations during the mar-riage."0 9 If valuation or compensation is not feasible, divorce will not occur,suggesting an important question: are these marriages that should continue?"'

Since we cannot test or compare parties' individual evaluations of their gainsor losses from their marriages,"' we cannot confirm Zelder's hypothesis thatdivorce in this setting involves greater losses to one party than the gains toanother."2 Without this information, the efficiency analysis collapses into adeep and troubling problem of the nature of marriage and love. This is a subjectthat Zelder, at least, is unwilling or unable to tackle.

Zelder points out, as have other economists, that mutual consent rules maybe seen as creating a property right in the continuation of marriage, whileunilateral no-fault rules create a property right in divorce." 3 Finding no eco-

108. That is, not valued enough to offset the other perceived losses from marriage, orperhaps not valued as highly by one partner as by the other.

109. See generally Margaret Jane Radin, Commodification and Commensurability, 43Duke L J 56 (1993); Margaret Jane Radin, Market Inalienability, 100 Harv L Rev 1848(1987); see also Estin, 36 Wm & Mary L Rev 1062-65 (cited in note 9). Cass Sunsteindescribes the anticommodification position as one which understands that people value"goods, things, relationships and states of affairs" not simply in different amounts, but indifferent ways. Cass R. Sunstein, Incommensurability and Valuation in Law, 92 Mich LRev 779, 782-85 (1994). To describe goods as incommensurable is to recognize that nosingle metric of value can adequately address the range of different valuations that peoplemake of those goods. Id at 795. In the context of marital bargaining, although mutualconsent proposals allow each party to make his or her own judgments as to the value (inquantitative terms) of marriage and divorce and all of the goods of family life, theproposals require that individuals treat these goods as commensurable. Moreover, it is hardto imagine what metric other than dollars the individuals could use in making these judg-ments.

110. Zelder does not consider the possibility that relatively lower divorce rates in thiscategory under fault rules may reflect the particular difficulties of negotiating divorce inthese circumstances. It is at least possible that the increase in divorce rates reflects a moveto a more "optimal" level.

111. In economics, of course, there is no methodology that permits interpersonal com-parisons of utility, or even measurement of utility, except in willingness to pay. See textaccompanying note 25.

112. Although Zelder suggests that he is applying a Kaldor-Hicks standard of efficiency,see Zelder, 22 J Legal Stud at 510 (cited in note 5), the premise of his model is that theparties act alone to assess their gains or losses from marriage and divorce, with no legalrules other than those creating a property right in marital status. This has much more incommon with a Pareto standard, which recognizes an allocation as efficient only when itis based on consent. A Kaldor-Hicks standard would allow an external assessment of thecosts and benefits to each party from divorce. See text accompanying notes 25-36 and 69.

113. Zelder, 22 J Legal Stud at 504 (cited in note 5); Zelder, 16 Harv J L & Pub Pol

[2:517

Economics and the Problem of Divorce 539

nomic distinction between enforced marriage and enforced divorce, he dismissesthe possibility that other values may be worthy of serious discussion." 4 Appar-ently, his preference for a property right in marriage is based solely on hisconclusion that this rule facilitates transferability of rights between husbands andwives. He does not address the arguments concerning the specification ofproperty rights,"' or the difference between property and liability rules.""

What is perhaps most remarkable is that this analysis claims to be a suffi-cient argument for sweeping and radical divorce policy reform. Although hecasually acknowledges that this approach to divorce might require a system to"regulate the allocation of resources within marriage,"" 7 he offers no consider-ation of what this regulation would mean for families or for law. Other econom-ic writers have confronted this problem, noting that mutual consent divorce isthe equivalent of orders for specific performance of marriage.s

b) Outcome models. Some proponents of mutual consent divorce movebeyond abstract bargaining models to a consideration of specific policy goals.Gary Becker, for one, argues that mutual consent divorce would reduce the roleof judges, "for couples would be required to work out their own terms as partof the consent process."" 9 In addition, he asserts that a change to mutual

at 246 (cited in note 5). See also Allen, 82 Am Econ Rev at 679 (cited in note 5).114. Zelder supplies this analysis in a footnote:[S]ome might quibble with the basic normative conclusion of this article, and notethat eliminating no-fault and thus, requiring fault, traps within marriage those ...who are made worse off by marriage. There are two responses to this contention.One is that if no-fault is the rule, not only do inefficient divorces occur, but somepeople are trapped in divorce, even though it makes them worse off. There is noreason to endorse one form of ensnarement over another. Second, if one accepts thisproblem . . . , one remedy is to regulate the allocation of resources within marriageso that ... [the partner who benefits] shares his gains with [the partner who loses](to the extent possible given that some of his gains are in the form of publicgoods) so that she is made better off during marriage (or at least loses less). Sucha practice would, of course, be inconsistent with the state's traditional protection offamily autonomy. Nevertheless, such a rule might be economically efficient if it pre-vented trapped spouses . . . from wasting resources or harming their better-offmates in order to induce divorce.

Zelder, 16 Harv J L & Pub Pol at 259 n 32.115. Zelder never discusses such questions as how a property right in marriage could

be specified in order to avoid negative externalities and overutilization of such rights. SeeTrebilcock, The Limits of Freedom of Contract at 10 (cited in note 60) (The goal is to"internalize as fully as possible to a property rights holder all the costs and benefitsassociated with utilization of the property rights in question."); see also text accompanyingnote 320.

116. Although Zelder notes the distinction between property rights and liability rulesdrawn by Calabresi and Melamed, he does not appear to consider their arguments in hisanalysis. Zelder, 16 Harv J L & Pub Pol at 246 n 11; see generally text accompanyingnotes 62-68.

117. Zelder, 16 Harv J L & Pub Pol at 259 n 32.118. See text accompanying notes 151-72.119. Becker, Bus Wk at 22 (Dec 7, 1992) (cited in note 4). Becker writes:Judges now have a large part in divorce proceedings, even though they may not

19951

540 Roundtable

consent would "greatly improve the bargaining power of the many married

persons, especially women, who are hurt by no-fault."'20 Professor Becker

argues that this would have important consequences: "women would be more

able to stay home for a spell after having children, if they wanted to, because

they would have much less reason to fear being left in a financial bind." 2 '

These two aspects of his proposal have a different normative ground: rather than

focusing on bargaining as a goal in itself, they seek to shift current distributions

of wealth after divorce to encourage certain choices within marriage. Although

the argument for private ordering reflects a preference for bargaining and Pareto

efficiency norms, Becker's other arguments expand on the exchange modelsignificantly.

The most surprising of these is the argument that mutual consent ruleswould improve women's bargaining power in divorce. Conventionally, economics

is not concerned with the endowments of power or wealth that parties bring to

a transaction. 22 Becker's argument implies, however, that these should beimportant concerns in the design of divorce rules, and further that there are often

important differences between the situations of the parties prior to their divorcenegotiations that the law should take into account. He has an implicit distribu-

tive goal: to shift a greater share of family financial resources to divorced women

and children. In the context of Professor Becker's recommendations for reform,this concern appears closely connected to the broader social problem caused by

single-parent households.23

Similarly, Becker's recommendation that new divorce rules would better

allow for the costs and risks of women's investment in childrearing reflects a

broader concern, which is made clearer by his positive economic analysis ofmarriage and childrearing. 24 Becker evidently believes that divorce law can

serve as a lever to shift the incentives and penalties that now discourage tradi-tional divisions of labor within marriage. This argument also moves beyond the

focus on efficient divorce to include a normative vision of marriage.These three arguments place Becker's recommendations somewhere between

the exchange model and the enforcement proposals discussed below. Although

have enough knowledge to fit custody and support provisions to individual circum-stances .... There is every reason to expect that privatizing divorce proceed-ings-taking the judiciary out of them-would make them work much better, justas privatizing other activities has brought improvement.

Id.120. Id. The particular scenario he develops is described in the text accompanying note

134.121. Id. "No-fault divorce laws discourage married women from leaving the workforce

for several years to care for their young children, because they realize that they will needgood jobs if their husbands ditch them." Id.

122. See, for example, Posner, The Economics of Justice at 79 (cited in note 29) ("The'interpersonal comparison of utilities' is anathema to the modern economist, and rightly so,because there is no metric for making such a comparison.").

123. See text accompanying note 136.124. See text accompanying notes 43-50.

[2:S17

Economics and the Problem of Divorce 541

his argument for reduction of the importance of "the judiciary" in divorcereflects a characteristic economic preference for private rather than collectivedecisionmaking, and a preference for Pareto rather than Kaldor-Hicks norms, itcontradicts his other two arguments, which suggest the need for more complexmodels of family regulation.

Allen Parkman argues that mutual consent divorce rules would further"efficiency, fairness and equality by recognizing all the costs of divorce."' s Inhis view, mutual consent rules would give the party who does not want todivorce "an incentive to ask for compensation for these costs as a basis foragreeing to the divorce,""'2 which would assure that divorces occur "only whenthe collective benefits exceeded the costs."'27 Moreover, "the dissolution ofmarriage would be based on the parties' criteria rather than those of thestate."' Parkman's argument for mutual consent rules, like Gary Becker's,describes the problem of divorce in distributive terms. He adopts the view thatfault-based divorce gave women a "veto power" that made it possible for themto negotiate better financial settlements.' Parkman also asserts that no-faultdivorce rules have discouraged traditional divisions of labor in marriage.'Although he also gives serious consideration to legal reforms that would imple-ment a contract enforcement approach to divorce, Parkman argues that themutual consent solution is preferable, because it permits a wife opposed to adivorce to demand compensation for all of the nonfinancial losses that resultfrom divorce.'

Becker and Parkman's recommendations imply that mutual consent rules canachieve efficiency goals in marriage and both distributive and efficiency goals indivorce.' The claim is very ambitious, perhaps because it is based on a rad-

125. Parkman, No-Fault Divorce at 123 (cited in note 6). Although Parkman advocatesa mutual consent system, id at 138, he would also allow a long-term separation as anadditional ground for divorce, id at 139, and permit no-fault divorce "during the first yearof marriage or until the wife becomes pregnant," id at 139-40.

126. Id at 138.127. Id at 137.128. Id at 140.129. Id at 44-46, 79-88, and 112.130. Id at 65-66, 101-03.131. Parkman states that a no-fault divorce rule would be the best (most efficient)

solution to the demand for marital freedom if there were "an appropriate definition ofproperty" which could be used to "divide the financial repercussions of the dissolution be-tween the spouses," and if divorce rules "recognize[d] the full costs of divorce." Id at 46;see also id at 63. He reasons that fault-based divorce rules with provisions for negotiatedsettlements are a "second best" solution. Id. Parkman views the divorce laws, as presentlystructured, as a third best alternative, concluding that they lead to inefficient divorcesbecause they do not "accurately reflect the reliance costs incurred by parties who do notwant the dissolutions." Id at 64; see also id at 111 ("The primary problem with no faultdivorce is that it ignores some costs of divorce."). Parkman concludes that mutual consentis better than no-fault because it is impossible to draft legal rules which account for allthe costs of divorce. Id at 122-23 and 137-39.

132. In his Business Week column, Professor Becker does not argue that mutual consent

1995]

542 Roundtable

ically oversimplified view of divorce. It seems designed to address the case inwhich a "married woman with young children cannot stop her husband fromdivorcing her" under the present rules.'33 The husband "very much wants a di-vorce and cares little about the harmful effects on his wife and children.3 4 Thegoal of mutual consent is to give her "leverage" to demand generous propertydivision and adequate assurances that he will visit and support the children.Certainly, this hypothetical mother is in a difficult situation, and the law shoulddo better by her.'35 As Becker implies, hers is a problem with significant conse-quences for the larger society, as well. 36 But there are many other compellingdivorce stories which may be less compatible with this type of regime.'37

Perhaps Professor Becker's proposal is best understood as an argument forprivatization, not just in the divorce process but in funding the social costs ofdivorce in families with young children. 3 This would be consistent with the

will reduce divorce rates to an "efficient" level but rather that it would permit one spouseto demand more substantial financial redistribution. Becker, Bus Wk at 22 (Dec 7, 1992)(cited in note 4). Other analysis suggests that the goals of private ordering and redistribu-tion may be incompatible. According to Professor Polinsky:

[I]t is frequently difficult, if not impossible, to use legal rules to redistribute incomein contractual disputes. . . . In general, the parties will take any distributionaleffects of breach of contract remedies into account when they negotiate the contractprice; thus, how the joint benefits are shared between the parties depends primarily,if not exclusively, on their relative bargaining strengths, not on the remedies avail-able to them.

Polinsky, An Introduction to Law & Economics at 122 (cited in note 78).133. Becker, Bus Wk at 22 (Dec 7, 1992).134. Id. Like Becker, Parkman describes divorce as though it were primarily a problem

with a single fact pattern, in which a husband with a promising career walks out on hiswife, who has specialized in household work. Parkman, No-Fault Divorce at 113-14, 122(cited in note 6); but see id at 142. This paradigm case pervades the literature; asProfessor Weisbrod points out, the discussion ordinarily focuses on "the remedies to begiven a spouse (typically assumed to be a wife) disadvantaged by the departure of a part-ner, [who is] often assumed to have the money but unwilling to give it up. The disadvan-tage is seen to arise from the fact that the wife either never seriously entered the marketor left it early." Weisbrod, 1994 Utah L Rev at 811 (cited in note 21).

135. See generally Ann Laquer Estin, Maintenance, Alimony and the Rehabilitation ofFamily Care, 71 NC L Rev 721 (1993).

136. Becker, Bus Wk at 22 (Dec 7, 1992) ("The plight of these divorced women andtheir children-who now constitute almost 20% of all households-is among the mostserious family problems in Britain, Canada, the U.S., and many other countries.").

137. What if the wife is seeking a divorce because of the husband's infidelity or herown? What if the parties have no children, but one partner has religious objections to di-vorce? What if one of the spouses is violent or abusive with the other? How generous aproperty division can the unwilling spouse demand? Everything? What if she in fact caresvery little about the children's welfare, and bargains only for a larger share of assets forherself? Is a mutual consent rule really fair or sufficient for all these circumstances?

138. This reading is consistent with other policy recommendations from ProfessorBecker, such as his suggestion that the use of private collection companies would increasethe rate of child support payment compliance. Gary S. Becker, Unleash the Bill Collectorson Deadbeat Dads, Bus Wk 18 (July 18, 1994). Both the objective of increased enforce-ment and the use of private collectors reflect a preference for non-governmental approaches

[2:517

Economics and the Problem of Divorce 543

trend toward greater private ordering in family law over the past two de-cades.'39 Seen in this light, a mutual consent approach might reduce the exter-nal costs of marriage and divorce in a number of ways. First, if it assurescustodial parents a greater share of family financial resources after divorce, itobviously reduces the need for public subsidy. Second, proponents of this viewexpect it would prevent some divorces altogether. 4 Third, the rules couldencourage parents, particularly mothers, to engage in household caregiving workrather than market labor."' This last point suggests a concern with a differenttype of externality, the wider social effects of increasing divorce rates. 42

Although the argument for privatizing the costs of divorce has obviousappeal, there is very little theoretical or empirical evidence that mutual consentdivorce laws can remove courts from the divorce process and divorced womenwith children from welfare rolls.'43 Whether or not this model works in theory,it is unrealistic to suppose that law can be so readily eliminated from theprocesses of marriage and divorce. Even in market settings, the process ofexchange is not self-executing or self-enforcing.' 44 At a minimum, some processmust control the risks of exploitation inevitably created by power differentialswithin marriage relationships. 45 The problem of unequal bargaining power,which Becker describes, might be more effectively addressed with better proce-dural rules and protections than with further privatization. Bargaining theoristsdemonstrate that the bargaining power of different parties is determined bybargaining endowments created by legal rules. 14 Similarly, Becker's normativeconcerns with marriage suggest the importance of an adequate set of background

to family problems.139. See Singer, 1992 Wis L Rev at 1444 (cited in note 16) ("In virtually all doctrinal

areas, private norm creation and private decision-making have supplanted state-imposedrules and structures for governing family-related behavior.").

140. Parkman argues that mutual consent divorce would "force parents to address thecosts incurred by their children due to a divorce," and he suggests that this could leadsome parents to choose not to divorce. Parkman, No-Fault Divorce at 123, 128 (cited innote 6).

141. See id at 78, 94-103.142. See text accompanying notes 423-28.143. There is growing academic debate on the question whether family issues of this

sort are adequately conceived as a matter of purely "private" responsibility. See, forexample, Deborah Rhode and Martha Minow, Reforming the Questions, Questioning theReforms: Feminist Perspectives on Divorce Law, in Sugarman and Kay, eds, DivorceReform at the Crossroads 191, 192-94 (cited in note 51); Harry D. Krause, Child SupportReassessed: Limits of Private Responsibility and the Public Interest, 24 Fam L Q 1 (1990).

144. See, for example, Trebilcock, The Limits of Freedom of Contract at 15-17 (citedin note 60); Robert H. Mnookin, Divorce. Bargaining: The Limits of Private Ordering, 18J L Reform 1015 (1985).

145. See generally Shelly Lundberg and Robert A. Pollak, Separate Spheres Bargainingand the Marriage Market, 101 J Pol Econ 988 (1993).

146. Mnookin, 18 J L Reform at 1024 (cited in note 144). As Mnookin points out, theobvious response to this problem is for the state to consider changes in the bargaining en-dowments of parties to divorce. This is what the "enforcement" approach attempts to do.See text accompanying notes 191-311.

1995]

544 Roundtable

legal norms against which private agreements could be tested. 4 7 Indeed, it isdifficult to imagine how private bargaining could proceed at all without legalrules that define and protect the entitlements which parties have available toexchange. 148

Another troubling aspect of the mutual consent proposals is that they would,by definition, force some unhappy individuals to remain in marriages againsttheir will. The set of legal and practical problems that this poses have not beenseriously discussed by authors in this area. 149 As several of these writers haveacknowledged, a mutual consent rule would shift the need for a public or judicialrole in family life from the point of divorce into the midst of family life it-self.'

2. Specific relief.

Of the advocates of mutual consent divorce, Allen Parkman most explicitlyrecognizes that it is the equivalent of a rule requiring "specific performance" ofmarriage contracts.' As Parkman notes, the ordinary remedial rules in con-tract dictate that specific relief is extraordinary, available only when monetarydamages would be inadequate to compensate for a breach.' In his view,

147. See text accompanying note 399.148. See, for example, Trebilcock, The Limits of Freedom of Contract at 44-48 (cited

in note 60); Robert H. Mnookin and Lewis Kornhauser, Bargaining in the Shadow of theLaw: The Case of Divorce, 88 Yale L J 950 (1979). Parkman sees the scope of thesenorms as critically important under a no-fault regime, but he treats them as relativelyinsignificant under fault-based or mutual consent divorce. Compare Parkman, No-FaultDivorce at 111-37 (cited in note 6), with id at 63 ("Under fault divorce, the legally de-fined financial arrangements served a minor role compared with the agreements workedout by the parties privately.").

149. For Zelder's assessment of the problem, see the quotation in note 114. Parkmanis almost equally cavalier. After he notes the problem of a spouse trapped in an unhappymarriage, he dismisses it with this observation: "None of the alternative approaches todivorce produces perfect outcomes." Parkman, No-Fault Divorce at 139 (cited in note 6).

150. In legal traditions, this sort of intervention is prevented by a range of rulesprotecting the privacy and autonomy of family life. See generally Estin, 36 Wm & MaryL Rev at 1039-53 (cited in note 9). Brinig and Crafton argue for greater attention toopportunistic behavior within marriage, and not simply at the time of divorce. Brinig andCrafton, 23 J Legal Stud at 883-92 (cited in note 5); see also text accompanying notes163-79.

151. He writes: "One of the major attractions of specific performance is its ability toforce the parties to estimate their costs-psychological as well as financial-due to thebreach." Parkman, No-Fault Divorce at 126-27 (cited in note 6). See also the quotationfrom Zelder in note 114.

Although Parkman recognizes some of the problems with a specific performance rule,he evidently concludes that these are less serious than the problems he identifies withunilateral no-fault divorce. See, for example, Parkman, No-Fault Divorce at 127, 138-140.At least in part, Parkman reaches this conclusion on the basis that a damages remedywould require courts rather than the individuals involved to measure the parties' intangiblelosses from divorce. Id at 127-28. Parkman does not explain, however, the basis for hisapparent belief that individual husbands and wives will be able to calculate these losses.See text accompanying notes 381-82.

152. Restatement (Second) of Contracts 5 359(1) (1981); see also Parkman, No-Fault

[2:517

Economics and the Problem of Divorce 545

however, this is the appropriate remedial analogue for divorce." 3 In our cul-ture, of course, husbands and wives are not generally viewed as fungible. Substi-tutes may be hard to acquire. As the situation of parties to a marriage is unique,ordinary damages remedies would be extremely difficult to compute.' 4

Another benefit of specific remedies in contract is that they can force partiesto negotiate a financial resolution, in which one party pays the other for arelease from the obligation to perform. 55 In theory, this leads to a sharing ofthe benefit from an efficient breach and increases the likelihood that the transac-tion will satisfy Pareto standards. s6 Parkman relies heavily on this aspect ofspecific performance in arguing for a mutual consent divorce rule; although hedescribes the rule as one which creates a right to continuation of the mar-riage,"57 he also asserts that "the courts would not normally become involvedin the quality of the performance during the marriage." 15

Divorce at 126. The finding of inadequacy usually requires that the promised performanceis in some way unique. If the performance is unique, it will be difficult to value,Restatement (Second) of Contracts S 360(a) and comment b. Conversely, if the perfor-mance is not unique, the nonbreaching party can be awarded damages to purchase a sub-stitute performance in- the market. Id S 360(b) and comment c; see also Posner, EconomicAnalysis of Law at 130-32 (cited in note 2) (damages are the only remedy unless thereis no good market substitute for the promised performance). As Judge Posner notes, thereis an extensive literature on the economics of specific performance. Id at 130 n 1.

153. Parkman makes the "uniqueness" argument in favor of the specific performance ap-proach to divorce. Parkman, No-Fault Divorce at 137.

154. Lloyd Cohen describes the measure as "the replacement cost of a spouse ofequivalent ex ante value." Cohen, 16 J Legal Stud at 298 (cited in note 6); see also textaccompanying notes 191-311.

155. As a number of contract theorists have argued, the issue has more to do with howthe gain from the breach will be shared between the parties. See, for example, Macneil,68 Va L Rev at 950-52 (cited in note 35). Macneil's argument for specific performanceremedies in the business setting stresses two important goals: encouraging a process ofconsultation prior to breach, and bringing all of the losses from breach into the parties'decision-making process. Id at 958-60. He describes the "simple efficient breach" theoryas biased in favor of uncooperative behavior (and litigation) rather than negotiationbetween the parties. Id at 968-69. On this front, however, the analogy between businessdeals and marriage relationships may not be very strong. Most divorces are not under-taken lightly, and in most marriages, including those already in deep trouble, there is acontinual process of communication, which is very different than what the simple breachscenario suggests.

156. The economic arguments are particularly compelling where the breach is op-portunistic. Posner, Economic Analysis of Law at 131 (With opportunistic breach, bydefinition, completion of the contract is not impossible or uneconomical.). Other theoristsargue that specific performance should be ordered where breach is "willful." See PatriciaMarschall, Willfulness: A Crucial Factor in Choosing Remedies for Breach of Contract, 24Ariz L Rev 733, 760-61 (1982).

Peter Linzer emphasizes the loss valuation aspect of the "uniqueness" test, arguingthat a specific performance remedy should be available in "any transaction in which thepromisee's damages cannot be ascertained by a market valuation. Peter Linzer, On theAmorality of Contract Remedies-Efficiency, Equity and the Second Restatement, 81Colum L Rev 111, 125 (1981).

157. Parkman, No-Fault Divorce at 126-27 (cited in note 6).158. Parkman, 8 BYU J Pub L at 95 n 11 (cited in note 6); see also Parkman, No-

1995]

546 Roundtable

But specific performance is not an order to negotiate damages: sometimes,perhaps commonly, the remedy means that the parties will be compelled toperform their contractual obligations. Because of the extraordinary coercion this

implies, contract law includes old and well-elaborated reasons for proceedingcautiously with specific remedies. Courts hesitate to order specific performancewhere it would require extensive supervision, or where it would order con-tinuation of performances viewed as personal. The economic literature alsoelaborates serious problems with specific relief, including its creation of abilateral monopoly with high transaction costs and the third party costs whencourts are required to supervise and monitor performance."s9 Parkman ac-knowledges that these problems apply in the divorce setting, but apparently hedoes not regard them as serious enough to counsel against specific relief.16

In an article published prior to Parkman's work, Lloyd Cohen criticizedmutual consent rules precisely because they are equivalent to a specific perfor-mance regime."" Professor Cohen observed that the law cannot make the par-ties perform the obligations of a marriage contract, since "many of the acts that

a spouse has implicitly contracted to perform cannot be specified nor theirperformance monitored." 62 Moreover, "marital duties are to be performed ina certain spirit, and no court can succeed in forcing an unwilling spouse toperform marital duties in a spirit of love and devotion."13

These concerns are familiar in the setting of commercial contracts. Becauseof problems of supervision and coercion, specific relief is not generally grantedwhere the terms of the contract are uncertain, 64 or where performance requiresthe promisor's cooperation. 6 ' This is particularly true where a promise in-volves a personal relationship. 6 Courts may attempt to compel performanceby enjoining the contract breacher from performing for another party, but strong

Fault Divorce at 139.159. See Posner, Economic Analysis of Law at 131 (cited in note 2).160. Parkman, No-Fault Divorce at 127, 139-40. Unfortunately, Parkman also does not

discuss the fact that Lloyd Cohen, in an article cited elsewhere in Parkman's book,concludes that these issues render a mutual consent/specific performance approach to di-vorce unworkable. See Cohen, 16 J Legal Stud at 300-01 (cited in note 6). See alsoWeisbrod, 1994 Utah L Rev at 812 (cited in note 21) (discussing the writ or restitutionof conjugal rights). Curiously, Parkman suggests that "strict enforcement" of maritalcontracts would not be desirable. Parkman, No-Fault Divorce at 63. Although this wouldseem to be necessary in his mutual consent regime, he apparently assumes that all coupleswill in fact reach agreements to end their marriages, dispensing with the problem ofenforcement.

161. Cohen, 16 J Legal Stud at 300-01.162. Id at 300.163. Id at 300. Peg Brinig and June Carbone argue that no-fault divorce "represented,

as much as anything else, rebellion against the propriety of specific performance of maritalobligations." Margaret F. Brinig and June Carbone, The Reliance Interest in Marriage andDivorce, 62 Tul L Rev 855, 883 (1988).

164. Restatement (Second) of Contracts S 362 (1979).165. Id S 366.166. Id S 367.

[2:S17

Economics and the Problem of Divorce 547

policy arguments counsel against this sort of restraint. 67

Cohen also observes that a mutual consent rule would only limit divorce, notbreach. 6' The possibility that one party will breach a marital contract withoutseeking a divorce raises serious administrative problems under a mutual consentrule."'9 The familiar scenarios of adultery, desertion, and cruelty from the eraof fault divorce have surely not disappeared.' 7 If the party in breach does notseek a divorce, the nonbreaching spouse is at a serious disadvantage. Althoughsuch misbehavior can destroy the value of the marriage for the wronged party,that partner likely will not be able to bargain for compensation.' Still worse,if the party in breach refuses to consent to divorce, the victim may be forced topay for freedom when the breach becomes sufficiently intolerable.'7

As noted above, where mutual consent is the only escape from an unhappymarriage, the risks of overreaching and other contracting failures become muchmore serious. Even in our present system, where much less of consequence rideson obtaining consent, divorce settlements often entail significant coercion. 3

For victims of marital battering, who should not have to purchase their freedomfrom violence, these are serious issues. 4 They also pose a problem for partiesto a marriage who do not have sufficient resources to purchase their partners'

167. Id § 357(2), 367(2), and comment c. This approach is reminiscent of the formerpractice in some states of granting divorces but prohibiting remarriage. See Clark, TheLaw of Domestic Relations 5 12.1 at 409 (cited in note 99).

168. Cohen, 16 J Legal Stud at 300-01 (cited in note 6). Parkman admits this as well.Parkman, No-Fault Divorce at 139 (cited in note 6) ("Mutual consent would not providea solution for the situation in which one spouse is the victim of acts such as cruelty oradultery, but the 'guilty' spouse does not want a divorce. Courts during the fault divorceera showed little skill, however, at making determinations in these cases .... Therefore,mutual consent would be no worse than fault divorce was in dealing with these situa-tions.").

169. Cohen, 16 J Legal Stud at 300-01. As Cohen argues, relaxed mores in our eraallow individuals to reap many of the gains of divorce without need for a formal decree.Id at 300.

170. Civil and criminal penalties for adultery have been significantly reduced over thepast twenty years. See Clark, Law of Domestic Relations at 498. Given this shift, thistype of misconduct may be more prevalent today.

171. Cohen describes this problem as the destruction of quasi rents generated by themarriage. See Cohen, 16 J Legal Stud at 301; see also text accompanying notes 205-16.Once there is a serious breach, often neither party wants the marriage to continue.Settlements worked out in these circumstances are "likely to be far different from whatwould have been specified ex ante as part of a liquidated-damage clause of the originalmarriage agreement." Id.

172. But see text accompanying notes 175 and 182.173. Sally Burnett Sharp, Fairness Standards and Separation Agreements: A Word of

Caution on Contractual Freedom, 132 U Pa L Rev 1399 (1984). See also Posner, TheEconomics of Justice at 190 n 35 (cited in note 29) ("The husband could make the wife'slife unbearable in an effort to force her to 'agree' to a divorce.").

174. See Brinig, 26 Fain L Q at 455 (cited in note 51); Linda J. Lacey, MandatoryMarriage "For the Sake of the Children": A Feminist Reply to Elizabeth Scott, 66 Tul LRev 1435, 1443-46 (1992) (describing the need for divorce when one spouse is abusive).

1995]

548 Roundtable

consent to a divorce.'" s Yet Parkman's proposal does not admit that judicialoversight of divorce bargaining might be appropriate, let alone necessary, toassure some level of fairness in results.' 76

Ultimately, the greatest failing of mutual consent proposals results from theformulation of divorce as an efficiency problem. The goal of achieving optimallevels of divorce places too much concern with divorce and too little concernwith marriage. Because of this narrow vision, the proposals never seriously noteor address the immense problems such a system could create on other fronts.

In order for a specific relief/mutual consent approach to work as writers likeParkman anticipate, the courts would have to take seriously an obligation toregulate both the negotiation and performance of marital and divorce contracts.Without such a framework, many of those the system is intended to protectwould be made much more vulnerable. But the difficulties of this project areimmense. Even in the commercial setting, specific remedies are extraordinaryrather than exclusive. Our legal system has never sought to assert this controlover family life, and a system that attempted to do so would conflict dramatical-ly with our norms of privacy and autonomy. In addition, the direct financialcosts of such a project would be enormous.

The mutual consent proposals regularly ignore these issues, largely becausethe economic models severely abstract the reality of family life. By deferringabsolutely to decisions based on the subjectively valued utility functions ofhusband and wife, these rules deny any broader social interest in ongoing familylife or divorce.' 77 The abstraction of divorce as exchange eliminates any under-standing of the contexts in which "bargaining" occurs. Yet context becomesmuch more important under rules which lead to enforced marriage.'78 And,because the efficiency model ignores market and contracting failures, it cannotaddress the myriad distortions of the fairness and efficiency of divorce bargains.As a result, these models do not provide for the effects of violence and spite,poverty or wealth, and strategic behavior of every variety."

175. Parkman concedes that it may be appropriate to loosen his scheme to allow.separation for a long period" as an additional ground for divorce. Parkman, No-FaultDivorce at 139 (cited in note 6).

176. Compare Trebilcock, The Limits of Freedom of Contract at 47-48 (cited in note60).

177. While proponents argue that mutual consent would give more protection andpower to women who now suffer financially in divorce, there is very little to suggest thatmost women could take advantage of their power to deny consent to achieve the resultsthese theorists imagine possible. See generally Penelope E. Bryan, Killing Us Softly: DivorceMediation and the Politics of Power, 40 Buff L Rev 441 (1992) (reviewing literature anddescribing problems with mediated divorce agreements).

178. None of these theorists suggests any inclination to begin regulating family behavior,however. See, for example, Zelder, 16 Harv J L & Pub Pol at 259 n 32 (cited in note5 and quoted in note 114) (noting that the regulation of the allocation of resources withinmarriage might be economically efficient but'nonetheless is inconsistent with the traditionalnotions of family autonomy).

179. These issues are much more central to a "transaction cost" analysis of marriageand divorce, which emphasizes the difficulties of "using contracts to structure complex,

[2:S17

Economics and the Problem of Divorce 549

The efficient divorce approach is also difficult to reconcile with the positiveeconomic explanation of increasing divorce rates. 80 These models do notexplain why the legal response to such realities as imperfect information, chang-ing tastes, and reduced gain from specialization should rigidify marital status andmake change more difficult. These issues are profound, and they are extremelyimportant in the formation of divorce policies. Paradoxically, these economists'recommendations contradict the economic paradigm that rational individualsmaximizing their own utility produce the best social outcomes.'' Thus, al-though the positive analysis of family behavior is founded oh rational choice, thenormative analysis apparently concludes that individually self-interested decisionsshould not remain the foundation of marriage and divorce.'

Economists are certainly not the only advocates of making divorce moredifficult, particularly for couples with children.' But these proposals seemsurprising when they emerge from a tradition which has always stressed theimportance of individual freedom to enter into and exit from relationships. In thenew economic theories of the family, it is a deep and often unacknowledgedproblem how far the paradigm of individual consent should extend. One can cer-tainly develop economic arguments as to why the protective functions of familylaw justify a more pervasive regulation of consent and exchange. These surelyinclude the significant third-party effects of divorce decisions on children andsociety, and market and contracting failures.'84 Economic theory offers a vo-cabulary and analysis that are useful in explaining some of these issues. But totoss aside the premise that each individual should be free to assess and act on hisor her own best interests is a radical break with the economic model, and onethat has not been well explained.' s

ongoing relationships." Pollak, 23 J Econ Lit at 595-96 (cited in note 5). In this ap-proach, bargaining between husband and wife extends not only to divorce decisions butalso to allocations within the family over long periods of time. This analysis of law isconcerned with its role in facilitating long-term relationships by reducing the possibilitiesfor exploitation and opportunism they create. See generally id at 581-608.

180. See text accompanying notes 37-45.181. This paradigm is the premise of the economists' move into positive analysis of non-

market behavior. See text accompanying notes 7-8. Given the central position of thisparadigm, some explanation of why it fades from view in these policy recommendationsseems essential.

182. It would be very useful to learn how Professor Becker combines his insights aboutthe various non-altruistic motivations of family behavior, see Becker, 101 J Pol Econ 385(cited in note 1), with his mutual consent approach to divorce. In his Business Weekarticle, he does suggest the possibility that a spouse would refuse consent out of spite andproposes dealing with that by "allowing a person to seek binding private divorce arbitra-tion if the spouse has not agreed to divorce terms after several years of separation."Becker, Bus Wk at 22 (Dec 7, 1992) (cited in note 4).

183. Among others, see Judith T. Younger, Light Thoughts and Night Thoughts on theAmerican Family, 76 Minn L Rev 891, 900-11 (1992); Elizabeth S. Scott, RationalDecision Making about Marriage and Divorce, 76 Va L Rev 9, 91 (1990).

184. See text accompanying notes 356-444.185. As Becker argues, economists have not fully developed methods for analysis of

1995]

550 Roundtable

The solutions proposed by writers like Becker and Parkman seem to be basedon a very narrow image of divorce: divorce is a problem caused by men withsubstantial resources who seek to escape from family responsibilities to marrynew wives.'86 This implies a correspondingly narrow view of marriage, basedon an ideal of the rational and efficient nuclear family with clearly establishedgender roles that reached its zenith in the golden age before no-fault divorce.1 7

Yet as Becker himself has described, family forms have evolved significantly inthe last few generations,' 8 and it is neither possible nor desirable to try to"turn back the clock."8 9 By hypothesis, new forms and structures of marriageare also rational; economic efficiency may dictate different divisions of labor inthe home. 98

B. CONTRACT ENFORCEMENT PROPOSALS

A different set of economic approaches to divorce places greater emphasis onthe norms of marriage and envisions a more significant role for law in facilitating

choices based on a wider spectrum of values than self-interest. See Becker, A Treatise onthe Family at 277-78 (cited in note 3) (noting the role of altruism in family life); Becker,101 J Pol Econ at 386, 398-400 (cited in note 1) (noting the role of anger, loyalty, spite,guilt, obligation, sense of duty, and masochism); see also text accompanying notes 463-64.The eclipse of consent in the writing on the family may result from the shadow cast byall of the varied factors, many of which, though potentially a basis for "rational" choice,are values we as a society prefer to suppress.

186. See text accompanying notes 133-37.187. In this respect, the economists' description of marriage and its virtues echoes the

functionalist sociology of the 1950s and 1960s, particularly the work of Talcott Parsons.See generally Talcott Parsons, The Kinship System of the Contemporary United States, inEssays in Sociological Theory 177, 177-96 (Free Press, revised ed 1954); Talcott Parsonsand Robert F. Bales, Family, Socialization and Interaction Process 3-26 (Free Press, 1955).Even the ideal of efficient, rational divorce is reflected here. See id at 19-20, 24-25.

188. Becker, A Treatise on the Family at 342-61 (cited in note 3).189. Becker, Bus Wk at 22 (Dec 7, 1992) (cited in note 4).190. It is increasingly difficult to sustain the traditional image of a family as neatly

divided between household and market producers. Many families now delegate manyaspects of child care, house cleaning, yard work, and other household services. MargaretBrinig has argued that it is not sensible to assume that all specialization occurs betweentwo members of a household without regard to third parties. Brinig, 26 Fain L Q at 457(cited in note 51); see also June Carbone and Margaret F. Brinig, Rethinking Marriage:Feminist Ideology, Economic Change, and Divorce Reform, 65 Tul L Rev 953, 990 n 168(1991); Brinig and Carbone, 62 Tul L Rev at 866 (cited in note 163). Following in theevolutionary tradition of social science analysis, it might be argued that the family is nowshedding its nurturing functions as it once cast off its economic and political roles.Parsonian sociology defended the process by which the family's functions were absorbedby the state and institutions such as schools, hospitals, and asylums. See Parsons andBales, Family, Socialization and Interaction Process at 3; Becker, A Treatise on the Familyat 342-61. For criticism, see Christopher Lasch, Haven in a Heartless World: The FamilyBeseiged (Basic Books, 1977); Eli Zaretsky, The Place of the Family in the Origins of theWelfare State, in Barrie Thorne and Marilyn Yalom, eds, Rethinking the Family: SomeFeminist Questions 188 (Longman, 1982).

[2:S17

Economics and the Problem of Divorce 551

family relations. These approaches also build directly on positive economicmodels of marriage, with particular attention to specialization of labor in thehousehold. Elisabeth Landes, in an article published in 1978, developed atheoretical bridge between this explanation of marriage and the entitlements ofdivorce."" Landes demonstrated that alimony rules can be understood as ameans to achieve efficient allocations of property rights upon divorce and that"the award and enforcement of alimony payments by the courts encourageoptimal resource allocation within marriage, increase the gain from marriage,and encourage the formation, productivity and stability of marriage."192

Landes's theory assumes that divorce bargaining does not occur under the idealconditions required for efficient outcomes.'93 In her analysis, another device,such as alimony, is necessary to promote efficiency both during the maritalrelationship and at its end.' 94

Since Landes's article, other writers have elaborated on the role that financialawards on divorce play in regulating marital behavior. The economic modelexplains divorce remedies with a set of analogies to commercial contract law.Law is understood to facilitate complex and long-term relationships in a varietyof ways: by providing "background" or "default" rules that control if parties failto negotiate all details of their relationship,19 by defining the parties' propertyrights and other legal entitlements as a condition for establishment of market orexchange relationships," 6 and by establishing mechanisms for enforcement orexcuse of contractual obligations." 7

191. Landes, 7 J Legal Stud 35 (cited in note 5).192. Id at 35-36. Landes's empirical analysis suggested that in the period prior to the

no-fault divorce reforms, alimony awards were "directly related to measures of the wife'shousehold specialization, such as children and (inversely) the wife's earning capacity, andto measures of the gain from marriage such as husband's income and duration of marriageprior to divorce." Id at 62.

193. She writes:[]f all marital income were perfectly divisible (i.e., not public goods) and if spousescould negotiate with each other and transfer income between themselves costlessly,a legal rule requiring mutual consent for divorce would be equivalent (in mostrespects) to one permitting unilateral divorce by either spouse.

In the absence of these conditions, alimony serves as an efficient means ofredistributing the property rights and assets of the marriage partnership between thespouses, enabling them to reach an "optimal" end-the dissolution of their marriage.

Id at 35.194. Id at 35-36.195. Trebilcock, The Limits of Freedom of Contract at 17 (cited in note 60). There is

an enormous literature on default rules in contract. See, for example, Symposium onDefault Rules and Contractual Consent, 3 Law & 1 (1993). Fully specified agreementsmay not be feasible for a wide range of reasons: high transaction costs, problems ofuncertainty and imperfect information, incapacity (for example, of infants), or problems ofexternalities. See generally Posner, Economic Analysis of Law at 92-93 (cited in note 2);Polinsky, Law and Economics at 29-32 (cited in note 78).

196. Trebilcock, The Limits of Freedom of Contract at 9-10 (cited in note 60); see alsoTrebilcock and Keshvani, 41 U Toronto L J at 551 (cited in note 6).

197. Trebilcock, The Limits of Freedom of Contract at 16.

1995]

552 Roundtable

This model is rooted in Kaldor-Hicks efficiency principles. It draws itsnormative force from a broad social cost-benefit analysis, rather than the actualconsent of the parties. Economists have attempted to link the analysis into Paretonorms, however, arguing that the background rights established by the legalsystem should "represent the entitlements that rational parties would havechosen ex ante (that is, at the time of marriage) had they negotiated a fullyspecified contingent claims contract." 98 Or, as Professor Becker argues, thestate regulation of divorce seems (or once seemed) designed "to mimic the termsof contracts between husbands and wives and parents and children that are notfeasible." 99 Moreover, they argue that the regime should be contractual, sothat parties may choose instead to negotiate their own terms in the place of thosesupplied by the law.2"'

Economists tend to assume that the terms a husband and wife wouldnegotiate are terms that maximize the return from their investment in marriage.Building on the positive economic analysis of families and households, this sug-gests that parties would prefer a relatively traditional household division oflabor.2"' This implies in turn that a husband and wife, negotiating at the outsetof their relationship, would choose terms that best make possible and protect thistype of arrangement.

Building on the contract metaphor, this model implies that marriage anddivorce law must include both substantive terms and a set of remedial provisions,analogous to rules governing damage awards in contract."2 The fact that rem-edies in contract are premised on breach has forced proponents of enforcementmodels to reconsider the merits of fault-based divorce. Many reach the trouble-some conclusion that some type of fault rule may need to be returned to thedivorce process in order to allow determination of liability for breach.20 3 Inaddition, the proposals devote substantial energy to the question of how "dam-ages" could be determined.2 4

198. Trebilcock and Keshvani, 41 U Toronto L J at 551. See also Cohen, 16 J LegalStud at 287-89 (cited in note 6). Cohen argues that such rules are justified, in divorce asin contract, because they are in the interest of both parties at the time the relationship isformed. Id.

199. Becker and Murphy, 31 J L & Econ at 14 (cited in note 6).200. Trebilcock, The Limits of Freedom of Contract at 47-48 ("In constructing an

appropriate background or default rule, I would, of course, allow parties to displace it byagreement if they so wished."). But see Regan, 1994 Utah L Rev at 645-47 (cited in note16) (arguing for a view of marriage as a social institution that imposes obligations that"cannot always be reconstructed as voluntarily assumed by self-interested actors").

201. See generally Becker, 101 J Pol Econ 385 (cited in note 1); see also text accompa-nying notes 233-37 and 243-45.

202. Carbone and Brinig describe this more broadly as a problem in "civil obligation."Carbone and Brinig, 65 Tul L Rev at 957-61 (cited in note 190); see also text accompa-nying notes 283-311. Other writers who advocate contractual approaches to regulatingmarriage appear less interested in wrestling with the issues of breach and remedy. See, forexample, Scott, 1994 Utah L Rev at 722 (cited in note 22).

203. See text accompanying notes 247-69. See, for example, Cohen, 16 J Legal Stud at303 (cited in note 6); but see Ellman, 77 Cal L Rev at 23-24 (cited in note 6).

204. See text accompanying notes 270-98. Brinig and Crafton point out that fault could

[2:517

Economics and the Problem of Divorce 553

1. Efficient marriage.

In the contract enforcement approach to divorce, the central concern is toencourage efficient behavior within marriage rather than to protect disadvan-taged spouses at the time of divorce."' That marital agreements are largelyindeterminate and unenforceable seems to discourage reliance on marriage andthus to undermine the goal of efficiency in marriage. Several writers havedescribed how traditional arrangements put wives at an economic disadvantageand have discussed methods to reduce the "risk and uncertainty of marriage,"and limit the parties' ability to engage in opportunistic behavior, to protect thesphere of household production.20 This is recognized as especially importantfor marriages which "produce" children.20 7

With the advent of the new home economics and household production

be used in setting alimony awards even if is not resurrected in the grounds for divorce,and of course, this is the practice in a number of states. Brinig and Crafton, 23 J LegalStud at 892-93 (cited in note 5). All of this is, of course, easier said than done, and mosteconomic and legal theorists recognize the enormity of the project of computing damages.

205. See Brinig and Crafton, 23 J Legal Stud at 892; Ellman, 77 Cal L Rev at 49-50.206. In particular, the approaches are based on these policy goals: encouraging marriage,

see, for example, Cohen, 16 J Legal Stud at 303; Cornell, 26 Farn L Q at 123-24 (citedin note 6), discouraging divorce, see Ellman, 77 Cal L Rev at 41 ("The first goal is toencourage the durability of the relationship."), encouraging economically efficient (wealthmaximizing) family organization and behavior, see id at 49-50, Landes, 7 J Legal Stud at49 (cited in note 5), Cohen, 16 J Legal Stud at 295, 303, encouraging investment inchildren, see Ellman, 77 Cal L Rev at 71-73, and encouraging spouses to invest in themarriage and each other, see generally Brinig and Crafton, 23 J Legal Stud 869.

207. See Estin, 36 Wm & Mary L Rev at 1002-04 (cited in note 9) (reviewing law andeconomics treatment of parent-child relations). According to Becker and Murphy, "[l]awspunish child abuse, the sale of children, and unauthorized abortions. They providecompulsory schooling, welfare payments to families with dependent children, stringent rulesabout divorce when young children are involved, and minimum ages of marriage." Beckerand Murphy, 31 J L & Econ at 1 (cited in note 6). See also Posner, Economic Analysisof Law at 149-50 (cited in note 2) (describing the role of the state in relation tochildren).

Further, economists posit that state involvement in the family substitutes for the contrac-tual arrangements children would make with their caretakers if they were cognitively andfinancially capable of entering into contracts. Becker and Murphy, 31 J L & Econ at 1,15. State intervention seeks to raise investment in children to socially efficient levels, andto offer adults in return the promise of support in old age. Id at 5-12. See generally LynnA. Stout, Some Thoughts on Poverty and Failure in the Market for Children's HumanCapital, 81 Georgetown L J 1945 (1993). See also text accompanying notes 430-39.Posner suggests that traditional marriage and divorce regulation reflect the importance of

child rearing, by permitting divorce in situations where "a husband's misconduct was likelyto hurt the children as well as the wife." Posner, Economic Analysis of Law at 145 (citedin note 2) (citing grounds of insanity, extreme cruelty, and criminality). Posner notesadultery as a different case, defending the double standard that traditionally punished awife's adultery more seriously on the basis that it protected husbands against the obliga-tion of rearing children that were not their own, and granting annulments where therewas fraud concerning a party's reproductive circumstances. Id at 141-45.

1995]

554 Roundtable

theory, economists began to explain marriage in terms of the specialization anddivision of labor between household and market activities.2 8 In theory, thisspecialization generates gains even absent the overlay of traditional gender roles,but economists are quick to note that those roles may be explained by women'scomparative advantage in production of children and perhaps other householdcommodities.2"9 Economists explain the risks that go with specialization inhousehold rather than market work in terms of investment in human capital. Ahomemaker invests in marriage-specific capital, which increases the productivityof the household but which is virtually worthless if the marriage dissolves." '

By contrast, a married person who works in the market sector develops humancapital which has value independent of marriage. This difference leaves ahomemaker particularly vulnerable in the face of divorce. Her economic alterna-tives to marriage are bleaker,' her opportunities for remarriage fewer,21 2

and, without financial transfers from her husband, she is left to bear the fullburden of long-term opportunity costs of household production.1 3

Changes in women's economic opportunities and in the social institutions ofmarriage have reduced the benefits of specialization within the family andincreased the homemaker's risks from divorce.2"4 Many economic analyses seekto stem this tide, by measures intended to protect women who have invested inmarriage-specific capital. 21 Given the asymmetric positions of husband andwife, many commentators are concerned particularly with exploitation andopportunism in divorce bargaining.216

208. See generally Posner, Economic Analysis of Law at 139-42; Becker, A Treatise onthe Family at 30-48 (cited in note 3).

209. Becker, A Treatise on the Family at 31-39.210. See Pollak, 23 J Econ Lit at 602 n 53 (cited in note 5), for a discussion of the

difference between marriage-specific and household-specific capital.211. This is especially true if her household work comes at the cost of developing skills

and experience that are valued outside a particular family. And, given her less attractivealternatives, she has correspondingly less bargaining power in negotiations over divorce. SeeLundberg and Pollak, 101 J Pol Econ 988 (cited in note 145).

212. Many economists are particularly concerned with this point. See text accompanyingnote 449.

213. See Estin, 71 NC L Rev at 746-48 (cited in note 135).214. See text accompanying notes 43-45.215. The literature on bargaining suggests that a housewife's alternatives are affected

during and after marriage by her specialization in household work. Allocations betweenhusband and wife depend not just on the total resources of the household but on theirindividual wealth, income, and earning power. Pollak, 23 J Econ Lit at 602; Lundberg andPollak, 101 J Pol Econ at 992. The housewife's work increases the gains from marriagefor both partners and decreases the gains she could expect from divorce. With lessattractive alternatives, she has less to bargain with at both stages. These effects are oftenamplified by the different experiences of women and men in the remarriage "market." SeePosner, Economic Analysis of Law at 148 (cited in note 2); Cohen, 16 J Legal Stud at288 (cited in note 6).

216. In Posner's description, "the fundamental function of contract law . . . is to deterpeople from behaving opportunistically toward their contracting parties, in order toencourage the optimal timing of economic activity and ... obviate costly self-protective

[2:517

Economics and the Problem of Divorce 555

The specialization that occurs during marriage is regularly compared to

specialization in business and employment contexts. 217 As Judge Posner pointsout, risks of layoff or termination in commercial settings are often compensatedby contract terms providing for severance pay, or for higher payments during theterm of employment. 1 He describes alimony as a type of compensation for ahomemaker in the event of a similar loss.

Taking a slightly different approach, Lloyd Cohen emphasizes that bothparties enjoy gains from their relationship. He frames the issue in terms of quasi-

rents generated during a contractual relationship,219 and points to the particularrisks of relationships in which the parties obtain benefits from the contract atdifferent stages of performance. " He points out that this long-term imbalanceprovides the opportunity for strategic behavior, most notably breach by the partywho has already received these gains." Professor Cohen argues that informalmethods'of protecting quasi-rents have become inadequate, making legal struc-tures more important.'m He concludes that protection of quasi-rents in mar-

measures." Posner, Economic Analysis of Law at 91.217. See id at 148; Cohen, 16 J Legal Stud at 287-89.218. Posner suggests that this explains the function of alimony payments:

Because the search for a suitable spouse is often protracted and because age maydepreciate a person's ability (especially if a woman) to form a new marriage thatwill yield her as much real income as the previous marriage did, it makes sense toinclude as a standard term in the implicit marriage contract a form of severance payor unemployment compensation that will maintain the divorced wife at her previousstandard of living during the search for a new husband.

Posner, Economic Analysis of Law at 148. Posner also writes:[I]n the case of marriage, the husband may be incapable of making the necessarytransfer payments to the wife, especially during the early years of the marriage,when the household may not have substantial liquid assets. Also, to calculate inadvance the appropriate compensation for a risk as difficult to quantify as that ofdivorce would be costly, especially since the relevant probability is really a scheduleof probabilities in each year of the marriage. This is a reason for awarding alimonyon a periodic basis even when it is awarded as a form of damages.

Id.219. Cohen defines quasi-rents as "a return to one party to a contract above what the

party would receive if the contract could be dissolved at will at that moment." Cohen, 16J Legal Stud at 287.

220. Cohen writes:As a rule, men tend to obtain gains early in the relationship when their owncontributions to the marriage are relatively low and that of their wives relativelygreat. Similarly, later on in marriage women tend as a general rule to obtain morefrom the contract than do men.

Id.221. Id. On the difference between appropriation and destruction of quasi-rents, see also

Trebilcock and Keshvani, 41 U Toronto L J at 556 (cited in note 6).222. See Cohen, 16 J Legal Stud at 289-91 (cited in note 6), for discussion of these

informal methods. Cohen argues that one gain from marriage is its insurance function,stated explicitly in traditional wedding vows. Id at 270, 288. Laws protecting quasi-rentsoffer another layer of insurance; this raises another economic issue, discussed as theproblem of moral hazard. See Trebilcock, The Limits of Freedom of Contract at 46-47(cited in note 60); Peters, 76 Am Econ Rev at 443-44 (cited in note 5).

19951

556 Roundtable

riage will promote greater investment in marriage.223

In one sense, this is positive economic analysis-a set of predictions as to thebehavioral effects that can be expected from a certain set of public policies. 224

But the analysis extends to a normative efficiency question: whether the behav-iors that result will either make the individuals affected feel they are better off,or increase the net social welfare. 22

' The writers primarily concerned withfostering efficient behavior in marriage generally do not argue that divorcegrounds should be limited to mutual consent.226 Rather, they recommendchanges in the rules that define the financial entitlements in divorce, to redistrib-ute the economic costs and benefits of specialization during marriage. 227 Theserules are intended to reverse present disincentives for spouses-particularlywives-to invest in family life. The argument runs that because this investmentis socially valuable and economically efficient, it should be encouraged, or atleast not discouraged, by the legal regime.228 In theory, however, rules that pro-vide these protections will also result in efficient decisions to divorce. 229

223. Cohen, 16 J Legal Stud at 296. Cohen argues that without legal or informalprotections in place, long-term investments in marriage have become a very risky venture.Economic analysis suggests that one response to increased risk would be a decreased levelof investment, and Cohen and others argue that this has occurred in the context ofmarriage. Id at 295. This is an area in which economists have attempted some empiricalproofs. See, for example, Parkman, 82 Am Econ Rev 671 (cited in note 5) (showing in-creased labor force participation rates for some married women in unilateral divorcestates); Peters, 76 Am Econ Rev at 451-52 (arguing that differences in divorce law areassociated with differences in labor force behavior but not fertility behavior).

224. Trebilcock, The Limits of Freedom of Contract at 3. Here, the assumption thatindividuals act as rational maximizers of their self interest leads to the corollary assump-tion that individuals will respond to these incentives. See id at 6; Posner, EconomicAnalysis of Law at 3-4 (cited in note 2).

225. Trebilcock, The Limits of Freedom of Contract at 7-8.226. Compare the discussion of Allen Parkman's arguments in note 131. Of those who

advocate a contract enforcement approach, only Lloyd Cohen deals expressly with mutualconsent divorce proposals. See generally Cohen, 16 J Legal Stud 267, and text accompany-ing notes 82-190.

227. This might be explained in terms of either property or liability rules. See textaccompanying notes 58-71. According to Trebilcock:

In defining and specifying property rights, an economic perspective would seekdefinitions and specifications that internalize as fully as possible to a property rightsholder all costs and benefits associated with utilization of the property rights inquestion. Failure to internalize costs may create negative externalities, leading toover-utilization of the resource in question from a social perspective. . . . Failure tointernalize benefits may create positive externalities, leading to under-utilization ofthe resource in question from a social perspective.

Trebilcock, The Limits of Freedom of Contract at 10.228. Some writers dispute the claim that traditional divisions of labor in the family are

efficient. See Jana B. Singer, Alimony and Efficiency: The Gendered Costs and Benefits ofthe Economic Justification for Alimony, 82 Georgetown L J 2423 (1994); Margaret F.Brinig, Comment on Jana Singer's Alimony and Efficiency, 82 Georgetown L J 2461(1994); Gillian K. Hadfield, Households at Work: Beyond Labor Market Policies to Rem-edy the Gender Gap, 82 Georgetown L J 89 (1993).

229. Cohen, 16 J Legal Stud at 296-303 (cited in note 6).

[2:517

Economics and the Problem of Divorce 557

The link between efficient marriage and efficient divorce requires thatdivorce bargaining be viewed within the larger perspective of marriage. Thus,Cohen and others argue that a system of divorce payments for homemakerswould satisfy efficiency norms by fostering behavior that is beneficial to bothparties when viewed from their perspective at the beginning of their mar-riage," whether or not the parties would otherwise agree to such compensa-tion at the time of divorce. Because of the enormous difficulty of negotiatingthese issues ex ante, the "agreement" is almost always hypothetical,"I but ifthe legal rules can be properly structured they will have the same positive effectson behavior. 2

The difficulty, of course, is that these hypothetical optimum terms must becollectively elaborated through a set of rules designed to structure the privatenegotiations around marriage and divorce, and to be applied by courts whereparties are unable to reach private agreement. 23 Moreover, it is essential thatpolicymakers "get it right," or they will create the wrong sorts of incentives. 4

In proposing specific default terms for the marital contract, economists tend toassume that efficiency, in the sense of achieving maximum output or production,is an uncontroversial goal within marriage. 235 This normative step is a precari-ous one, especially given the need to postulate hypothetical consent to these

230. Id at 303.231. Cohen concludes that "market alternatives," such as prenuptial contracts, are not

"an effective means of protecting spouses from inefficient breaches of the marriagecontract." Id at 268, 297-99. This is not to suggest that Cohen or other writers woulddeny enforcement of prenuptial agreements, but rather that they see fully specified long-term marital agreements as impractical. See Trebilcock and Keshvani, 41 U Toronto L Jat 556 (cited in note 6); Pollak, 23 J Econ Lit at 596 (cited in note 5) (Agreements are"costly or impossible to write, a reflection of bounded rationality and asymmetric informa-tion."). See text accompanying notes 362-82.

232. Carl Schneider raises the issue whether in fact divorce law has much effect onmarital behavior particularly when compared to short-term incentives and disincentives forparticular decisions. See Carl E. Schneider, Retbinking Alimony: Marital Decisions andMoral Discourse, 1991 BYU L Rev 197, 204-07, 211-14, 216; see also Ellman, 1991 BYUL Rev at 266-68 (cited in note 82). There is empirical evidence that individuals areunrealistically optimistic about the likelihood their own marriages will succeed. See LynnA. Baker and Robert E. Emery, When Every Relationsbip Is above Average: Perceptionsand Expectations of Divorce at the Time of Marriage, 17 L & Human Beh 439 (1993).

233. As Trebilcock notes, neoclassical economics is typically skeptical of collectivedecisionmaking, in part because of a failure to recognize that state elaboration of"background legal entitlements" is a prerequisite for private bargaining. Trebilcock, TheLimits of Freedom of Contract at 2-8, 15-17, 44, 56-57 (cited in note 60).

234. The literature notes the risk that generous alimony rules may create incentives foreconomic dependence. See, for example, Trebilcock and Keshvani, 41 U Toronto L J at539 (cited in note 6). This is typically described, in insurance terms, as a problem ofmoral hazard. See id at 557; Ellman, 77 Cal L Rev at 3 (cited in note 6).

235. Based on the assumption that the parties act rationally, and that rational actorswould seek to maximize their utility, the argument asserts that the hypothetical mar-riage/divorce contract terms should be those which lead to the highest levels of productionin the household. See Cohen, 16 J Legal Stud at 268-71 (cited in note 6).

1995]

558 Roundtable

terms.Whatever the explanatory power of positive economic analysis of family life,

it is an entirely different matter to claim that this model provides a sufficientnormative basis for legal regulation of family life. The claim that individualswould choose marriage contract terms in order to maximize their returns isextremely difficult to evaluate and probably impossible to quantify.236 This istrue even though "utility," while understood broadly in theory, is usuallyreduced to a readily measurable financial core in practice. 37

Transaction cost economists defend the rules around marriage and divorcesomewhat differently. Starting from the premise that it is not sensible to imaginethat parties negotiate at the outset about all aspects of their family or householdrelationships, '38 Douglas Allen argues that all varieties of state involvement inmarriage are means of avoiding the large transaction costs required to enter intoand enforce private agreements concerning marriage. 39 In this literature, mar-riage is viewed "as a governance structure, which permits some flexibility whileprotecting the parties against the hazards of unconstrained bilateral bargain-ing. " "' In effect, the law creates a standardized contract, saving significanttransaction costs, particularly when the costs of enforcement are taken intoaccount.24" ' This approach avoids the need to argue from hypothetical consents,but it maintains the central normative role of efficiency principles

Viewing marriage as a type of long-term contract, the losses from divorcecan be substantial. But developing a means for legal protection against theselosses has proven to be a complex and difficult project. 42 To the extent thatunilateral no-fault divorce permits an individual to end his or her marriagewithout compensation of a spouse's losses, it does not sort efficient from

236. As Pollak notes, "[b]ecause of the central role of unobservable variables (e.g.,preferences, household technology, genetic endowments), the new home economics view ofthe family does not lead simply or directly to a model capable of empirical implementa-tion." Pollak, 23 J Econ Lit at 584 (cited in note 5) (citations omitted); see also RobertA. Pollak and Michael L. Wachter, The Relevance of the Household Production Functionand Its Implications for the Allocation of Time, 83 J Pol Econ 255-77 (1975).

237. See Estin, 36 Wm & Mary L Rev at 1001-02, 1062-66, 1086 (cited in note 9),for a more general discussion of this phenomenon.

238. Robert Pollak contends that fully specified, long-term marital agreements are "costlyor impossible to write, a reflection of bounded rationality and asymmetric information."Pollak, 23 J Econ Lit at 596.

239. Allen, 13 J Econ Beh & Org 171 (cited in note 5).240. Pollak, 23 J Econ Lit at 603; see also Allen, 13 J Econ Beh at 173-75.241. Allen, 13 J Econ Beh & Org at 176. Allen argues that detailed private agreements

"are costly because each partner tries to bias the specific rules in his or her favor." Id at177-78.

242. Cohen's theory starts from the premise that as social controls on divorce haveabated, legal mechanisms have become more important to protect against the appropriationor destruction of quasi rents. As Cohen points out, the costs associated with breach havedeclined as divorce becomes more socially acceptable. At the same time, these changeshave undermined the informal enforcement mechanisms that once helped prevent breach.Cohen, 16 J Legal Stud at 289-90 (cited in note 6).

[2:517

Economics and the Problem of Divorce 559

inefficient transactions. But, by the same token, a rule forbidding divorce orrequiring mutual consent would prevent divorce in many circumstances in whichit would be efficient. Proponents of a contract enforcement regime seek thesolution to this problem in a framework for determining breach and damages.

2. Breach and damages.

Outside the realm of theory, regulating marriage and divorce to increasehousehold efficiency is an enormously complex undertaking. Economic theoryhas barely begun to specify the commodities that are produced within the family,the range of relevant inputs, and the mechanisms by which these are distributedamong family members, and it seems unimaginably beyond the realm of econom-ic expertise to expect a blueprint or even a menu of choices for the ideal efficientfamily.243 Given that current writers are not concerned with the specifics ofindividual preferences, developing terms for the hypothetical marriage contractwould require drastic changes in technique. Yet without more detail along theselines, how can policymakers determine what it is they are seeking to opti-mize?2' The choices seem endless, and they are constantly shifting in the faceof varying opportunities and life circumstances.24s

To the extent its terms are elaborated in law, the marriage contract is ageneral agreement, which recognizes a wide variety of legitimate preferencesconcerning such questions as the bearing and raising of children, attachment tothe labor force, or moral and spiritual commitments. This means that theincentives and protections in the law will be less than perfect. For efficiencyarguments, there are two critical components of contract-based regulation: rulesthat define when a breach may be said to have occurred, and rules that define

243. The difficulty of making these theories concrete is suggested by the high level ofabstraction used in this literature. See, for example, Becker, A Treatise on the Family at23-25 (cited in note 3); Allen, 13 J Econ Beh & Org at 176 (cited in note 5); Pollak, 23J Econ Lit at 598 (cited in note 5). For an example of the effort required to quantifythese fundamental issues, see Andrea H. Beller and John W. Graham, Small Change: TheEconomics of Child Support (Yale, 1993).

244. Marital duration? Fertility? Labor market participation? Autonomy? Love? Some of

the many questions will be philosophical: how much of the gain from family life ismaterial, and how much spiritual? What common currency can family members use tomeasure these benefits and costs? How do the different wants of different family memberscompare? What should be the effect of changes over the life course? Other questions aremore pragmatic: should the hypothetical marriage contract protect a new mother wholeaves the paid labor force entirely for fifteen years, or only six years, or only two years?And, for how many children? Should the contract promote or penalize a midlife careerchange off the fast track? Should the homemaking partner in a childless couple have thesame financial protection in the event of a divorce as one who cared for children?

245. See, for example, Pollak, 23 J Econ Lit at 601. Allen argues that "the state hasa comparative advantage in delineating rights only to the extent that marriages are fairlystandard, a condition met in the hierarchical and patriarchal marriages of the past." Allen,13 J Econ Beh & Org at 180 (cited in note 5). Certainly there are many reoccurring

patterns in marriage and divorce, but it is hard to argue that marriages are still "fairlystandard."

1995]

560 Roundtable

the remedies available in that event.246

Determinations of breach are difficult when the duties spouses owe eachother have not been clearly specified. 247 Before the no-fault reforms, a findingof marital fault stood somewhat in the place of a finding of breach in contractadjudication. 248 However, the only types of fault judicially and legislativelyrecognized as grounds were quite substantial. Even adultery, cruelty, or desertionwere not grounds for divorce unless they were sufficiently severe. 249 In a systemwhich depends on divorce rules to create incentives for economically efficientbehavior in marriage, this list appears to be radically incomplete.

Moreover, traditional fault rules would only have the effects economists seekif there is a correlation between economic losses from specialization and thelikelihood of a spouse committing substantial acts of adultery, cruelty, ordesertion. The only connection made between these two traits that has beenarticulated is based on gender, with the assertion that women are more likelyboth to have economic losses from divorce and to be the victims of spousalmisconduct.20 This connection is problematic on a number of grounds, and itis especially hard to accept as a basis for divorce policy without better empiricalevidence than the literature offers.2 s'

Margaret Brinig and June Carbone argue that the benefits of fault determina-

246. See, for example, Restatement (Second) of Contracts S 33(2) (In order for abargain to be enforced as a contract, terms must be reasonably certain. Terms "arereasonably certain if they provide a basis for determining existence of a breach and forgiving an appropriate remedy.").

247. See Cohen, 16 J Legal Stud at 298, 300 (cited in note 6).248. Ellman, 77 Cal L Rev at 23-24 (cited in note 6).249. Clark, The Law of Domestic Relations S5 13.2-13.4 at 497-510 (cited in note 99).

Many states also allowed divorce on grounds that do not correspond well to the idea ofbreach. See id at 510-12 (incompatibility) and 517-21 (living separate and apart).

250. It is routine in this literature to address divorce as a problem that primarily harmswomen. See, for example, Trebilcock, The Limits of Freedom of Contract at 44-46 (citedin note 60); Becker, 101 J Pol Econ 385 (cited in note 1); Cohen, 16 J Legal Stud at277. Among the reasons given are the greater likelihood that women will invest in maritalrather than market capital, Trebilcock and Keshvani, 41 U Toronto L J at 553-58 (citedin note 6), and the greater likelihood that women will lose value in the remarriage marketas a function of time, Cohen, 16 J Legal Stud at 273.

251. The problem with equating fault grounds with the causes of marriage breakdownwere debated at length at the time no-fault divorce reforms were adopted. Clark, The Lawof Domestic Relations S 12.1 at 410-12 (cited in note 99); but see Harvey L. Golden andJ. Michael Taylor, Misconduct as a Viable Consideration in Marital Dissolution, inAlimony: New Strategies for Pursuit and Defense (ABA, 1988) (providing arguments infavor of returning fault notions to divorce law.) For the argument that traditional fault-based divorce was not generally beneficial for women, see Jana B. Singer, Divorce Reformand Gender Justice, 67 NC L Rev 1103 (1989). Viewed historically, one problem was thesexual double standard in the treatment of adultery. Id at 1110-11; Cohen, 16 J LegalStud at 274-75, 278-87 (cited in note 6). A more contemporary concern is the difficultyof reconciling protections for women in economically dependent relationships with idealsof gender fairness and neutrality. See Trebilcock and Keshvani, 41 U Toronto L J at 535-37 (cited in note 6).

[2:517

Economics and the Problem of Divorce 561

tions "are those traditionally associated with civil obligation-deterring breachand encouraging reliance,"2 and they suggest that current rules "discouragerather than protect economic reliance on marriage.""' Although they recognizethe potential benefits of breach rules, their work also identifies significantproblems with returning fault concepts to the divorce process. 4 ProfessorsBrinig and Carbone point out that fault rules and an expectation measure ofdamages served in the past to protect particular reliance interests in marriage,namely the lost opportunity to marry differently." They point out, however,that while this type of loss may still be a factor, particularly for women, relianceinterests in marriage are now more varied, with the loss of career opportunitiestaking on relatively much greater importance. 6 This shift is significant becauseto the extent that legal protections for these different types of interests diverge,fault-based divorce grounds will have limited usefulness in protecting againsteconomic lossesY 7

In a system premised on findings of traditional marital fault, such as adul-tery, cruelty, or desertion, there is no general protection against the risks ofdivorce. Financial remedies are only available to a spouse if her partner has

252. Carbone and Brinig, 65 Tul L Rev at 959 (cited in note 190). These authors havedeveloped their analysis of divorce remedies in a number of articles, written together andseparately. See also Brinig and Crafton, 23 J Legal Stud 869 (cited in note 5); Brinig, 26Fam L Q 453 (cited in note 51); Brinig and Carbone, 62 Tul L Rev 855 (cited in note163); Carbone, 43 Vand L Rev 1463 (cited in note 36).

253. Carbone and Brinig, 65 Tul L Rev at 987. Their work draws distinctions betweenexpectation-, reliance-, and restitution-based theories of recovery and highlights the policychoices implied in each. They characterize compensation for opportunity losses that resultfrom marriage as a reliance measure of recovery, and they note that where once thereliance loss in divorce was primarily the lost opportunity to "marry well," it now consistsmore significantly of lost career opportunities. Brinig and Carbone, 62 Tul L Rev at 870-82, 894-96.

254. Brinig and Carbone, 62 Tul L Rev at 894-98.255. They argue that traditional alimony payments served as expectation damages, giving

the nonbreaching wife the financial equivalent of the support she would have enjoyed hadthe marriage continued for her lifetime. Carbone, 43 Vand L Rev at 1474-75. In effect,fault rules were the basis on which the law chose between men's and women's competingclaims at the time of a divorce. Brinig and Carbone, 62 Tul L Rev at 875.

256. They write:[D]ue to the expansion of employment opportunities for women, wives are lesslikely to be completely dependent on their husbands' income for support upondivorce. With women contributing to the family income, the husband's ability to re-marry no longer depends to the same degree on his [post-divorce] financial resourc-es. Moreover, for both men and women, the expectation value of marriage (and ofany foregone opportunities to marry) must be discounted by higher rates of divorce.Accordingly, while many women continue to rely on marriage to secure a standardof living greater than they could provide on their own, and many men continue torely on wealth in securing more favorable marriages, the value of marriage is lessthan in societies where divorce is rare, and divorcing couples are better able tomitigate their losses than they were in more traditional societies.

Brinig and Carbone, 62 Tul L Rev at 876-77 (cited in note 163) (citations omitted).257. Id at 895-96 ("A fault based system cannot adequately protect lost career op-

portunities or other contributions to their marriage.").

1995]

562 Roundtable

violated a set of behavioral norms which have almost nothing to do with thefactors that have put her at economic risk. From this perspective, advocating thereturn of traditional fault grounds is disconcertingly close to recommending thatthe entire package of traditional gender roles within marriage should be not onlyencouraged but enforced. 2

' By limiting the circumstances in which a home-maker spouse (read: wife) could hope to recover any financial award in divorce,the law would effectively make her responsible to insure the happiness of theparties' life together. She would have no recourse for an incompatibility short ofviolence, any disagreement over allocation of spending or leisure time short ofcomplete desertion, or any loss of love short of provable adultery by her spouse.To provide incentives for optimal behavior within marriage, these norms are notadequate. s9

Adjudication of fault in divorce also imposes substantial costs on the partiesand the system. 2" Divorce cases already represent a substantial portion of thecivil litigation in state courts,261 and in most of these cases one or both parties

258. Professors Carbone and Brinig refer to those who wish to retain or return faultconcepts to divorce as "traditionalists," who favor differentiated roles for men and womenin marriage. This category includes Becker, Landes, and in some respects Cohen. Carboneand Brinig, 65 Tul L Rev at 988-90 (cited in note 190).

259. Brinig and Crafton, 23 J Legal Stud at 887-88 (cited in note 5). Margaret Brinigand Steven Crafton argue that opportunistic behavior during marriage, including spousalviolence, is a more significant problem under no-fault divorce. They argue that even ifmarriages remained terminable at will, there should be financial remedies available forsignificant breaches that occur during the marriage. Id at 880-81. Without these remedies,they argue that marriage is an "illusory contract," and they offer empirical evidencesuggesting that unilateral no-fault divorce has had negative effects on behavior within mar-riage. Id at 883-92. As they put it: "[t]he puzzle becomes how to provide incentives sothat once again the majority of behavior clusters around what most people conceive of as'marriage.'" Id at 892.

On the question of how the terms of marriage contracts would be determined,Professors Brinig and Crafton suggest this requires looking to the "expectations of theparties," and they argue that the words used in most marriage ceremonies give somecontent to their agreement:

At the very least they mean that the parties intend their relationship to lastpermanently, that they plan to live together, care for each other, and give eachother support. They anticipate changes and adversity and pledge to work throughthem. In some ceremonies, the couple also make promises to "love and cherish" andto "forsake all others." Whether or not these words are used, the state will assumethat the spouses will be civil to each other, respecting bodily integrity. In moststates, there is the further assumption that the parties will be sexually faithful toeach other.

Id at 873-74 (citations omitted). See also Cohen, 16 J Legal Stud at 272-73 (cited in note6).

260. Although Brinig and Carbone recognize this point, they advocate a dual systemthat permits findings of fault "only after termination of a relatively long term marriage,"if there is a showing of "conduct sufficiently egregious to be clearly responsible for thedivorce." Brinig and Carbone, 62 Tul L Rev at 897 (cited in note 163) (citations omitted).In economic terms, this includes both transaction costs that are borne by the parties (time,legal and other fees, stress, humiliation) and costs that are externalized and subsidized inpublic funding for the legal system.

261. In 1992, domestic relations cases were a third of all civil filings in state courts of

[2:517

Economics and the Problem of Divorce 563

have no counsel.262 The expense and difficulty of these proceedings under fault-based divorce rules was one of the central reasons that they were abandoned. 263

Even in a dual system, in which the law permitted but did not require findingsof fault in divorce, many of these problems would continue. 2

6 After discussingthe benefits and drawbacks of several legal approaches to protecting quasi-rentsin divorce and property settlements, Professor Cohen concludes that none ofthese approaches is satisfactory. Recognizing the drawbacks of a system in whichcourts award compensation to the nonbreaching party, he states that only theshortcomings of the other approaches make this alternative plausible.26

1

The debate over rules for breach puts in tension the difficulties involved indefining and enforcing codes of marital behavior on the one hand and finding aprincipled basis for financial liability without fault on the other.266 Some com-mentators do not advocate restoring fault. Ira Ellman has developed a theory ofalimony to allow more room for private ordering of family relationships withoutreturning to a regime based on fault. Professor Ellman rejects the "contractanalogy" in divorce, largely because he sees determinations of breach as incon-sistent with the no-fault reforms.267 Professor Ellman argues that there are no

general jurisdiction across the United States. Brian J. Ostrum, et al, State Court CaseloadStatistics Annual Report 1992 23 (Nad Ctr for State Courts, 1994). The National Centerfor State Courts reports that there has been a substantial increase in this portion of thecourts' caseload since 1988. Id at 23-24.

262. See note 413.263. Herbert Jacob, The Silent Revolution: The Transformation of Divorce Law in the

United States 91-92 (Chicago, 1988). Allen Parkman recognizes this point when he de-scribes fault divorce laws as a "second best" solution to the demand for easier divorce.Parkman, No-Fault Divorce at 63-64 (cited in note 6).

264. Richard Neely, The Divorce Decision 44-53 (McGraw-Hill, 1984) (providing hypo-thetical illustrating practical difficulties of mixed fault/nonfault divorce grounds).

Brinig and Carbone also point out that the computation of a spouse's losses cannotbe mathematically precise and therefore must be accompanied by an inquiry into thespouse's needs and the partner's ability to pay. Brinig and Carbone, 62 Tulane L Rev at901-03.

265. Cohen, 16 J Legal Stud at 299-303 (cited in note 6). The approaches he discussesare unilateral divorce, mutual consent divorce, indissoluble marriage, and court-determineddivorce settlements. Id. In a letter to me, Cohen states that he intended to write a piecefavoring mutual consent divorce, but that as he wrote it he "came to see the importantways in which this contract was different from commercial contracts and therefore whymutual consent would do a very poor job of solving this problem of compensating thevictim of a breach of contract." He describes his position in the piece as "that of thenihilist who criticized all the legal alternatives" and that he sees the article as "a de-pressing conservative paper in which both economics and the law were helpless to makethings work out well." Letter from Cohen to Estin (cited in note 74).

266. There is now an extensive literature on the problem of justification of alimony innon-fault divorce systems. See, for example, Mary E. O'Connell, Alimony After No-Fault:A Practice in Search of a Theory, 23 New Eng L Rev 437 (1988). Various approachessuggested include a model based on strict liability in tort, see Twila L. Perry, No-FaultDivorce and Liability without Fault: Can Family Law Learn from Torts?, 52 Ohio St LJ 55 (1991), and income-splitting methods, see Singer, 82 Georgetown L J at 2454-60(cited in note 228).

267. Ellman, 77 Cal L Rev at 16-24 (cited in note 6). Professor Ellman points out that

1995]

564 Roundtable

longer the shared social and economic understandings about the nature ofmarriage which are necessary for contract-based regulation of marriage. 2

" ForEllman, divorce entitlements should therefore be defined to compensate forcertain types of economic losses from marriage, independent of the question ofbreach or fault, rather than to provide a remedy for all losses from a failed mar-riage.

Building on the theories developed by Elisabeth Landes and Lloyd Cohen,Professor Ellman argues that alimony awards should be redesigned to compen-sate losses in earning capacity that result from efficient allocations of rolesduring a marriage.270 Professor Ellman describes his method as one which"emulates a conventional economic analysis of law" by asking "what wouldhappen if we had no alimony remedy at all, but instead left the losses where theyfell at the time of divorce." '27 1 He argues that without alimony, there are finan-cial disincentives for allocations of family roles that might otherwise be optimal.His argument is that the legal environment should be modified to eliminate"distorting incentives," with the goal of maximizing the parties' freedom to"shape their marriage in accordance with their nonfinancial preferences."2 72

For Ellman, the law should be neutral; that is, it also should not reward anyparticular set of marital roles.2 73 He believes this is most likely if the law pro-tects the parties against the risk that one spouse would suffer a much greater fi-nancial loss than the other in the event of divorce.274

Although Professor Ellman articulates as a policy goal encouragement of"the durability of the [marital] relationship, ' 27

' he clearly is more concernedwith encouraging optimal marital roles than with restricting divorce.2 76 He uses

reasoning in divorce cases from fault jurisdictions "often has a contract flavor." Id at 23.He argues, however, that "contract" is only a metaphor in family law. Id at 13-14. In hisview, use of contract remedies requires that terms of marital agreements be defined withsome specificity, id at 18-23, and that breaches be identifiable, id at 17-19. Ellman impliesthat he views the move away from fault as a wise one. Id at 7, 24; Ellman, 1991 BYUL Rev at 304 (cited in note 82); but see Schneider, 1991 BYU L Rev at 250-54 (cited innote 232) (arguing for incorporating moral concerns in no-fault divorce).

268. He rejects restitution and partnership analogies on the same basis, see Ellman, 77Cal L Rev at 13-24, and also argues that there are no external performance standards ofthe sort that operate in commercial contract settings, id at 28-29.

269. See id at 53-73.270. Id at 41 n 128, 42 n 130.271. Ellman, 1991 BYU L Rev at 264.272. Ellman, 77 Cal L Rev at 51.273. Ellman, 1991 BYU L Rev at 265. As he notes, where creation of positive incen-

tives is involved, there are additional problems of moral hazard and adverse selection. Idat 293 n 69. Because of his effort to avoid creating incentives, he rejects ProfessorCarbone's characterization of his proposal as one designed to channel women intohousehold work. Id at 288 n 63.

274. Ellman, 77 Cal L Rev at 50-51 (cited in note 6).275. Id at 41.276. See id at 56 ("Some marriages should be dissolved."). Ellman asserts that some

divorces that now occur would not take place if laws reallocated the parties' financiallosses. Id at 50. By the same token, however, other divorces might occur under a new

[2:517

Economics and the Problem of Divorce 565

efficiency norms to ensure the highest total income for the household.2" Hisdisclaimer that his theory would create positive incentives for particular arrange-ments suggests that he believes that husband and wife will naturally agree onefficient arrangements once the distorting influence of divorce laws is sweptaway.278 Professor Ellman defines very specifically the losses that he believes

should be recognized and adjusted upon divorce in the obvious belief that findinga balance between encouraging and discouraging certain behavior is complex andimportant.279

Professor Ellman builds his model on the assumption that husbands andwives behave as rational, self-interested bargainers and that they will avoidarrangements that entail serious financial risks.20 He advocates alimony rulesthat provide compensation for economic losses from "economically rationalsharing behavior,"21

' and specifically rejects the prospect of alimony based ona wife's long-term financial "expectation" from marriage.2

The homemaker's financial losses from specialization in the household are

system that would not now take place.277. Id at 46-47. Ellman's proposal has been criticized on the basis that it only

optimizes money income. See, for example, Carbone, 43 Vand L Rev at 1468-69 (cited innote 36); Schneider, 1991 BYU L Rev at 235-43 (cited in note 232); see generally Cornell,26 Fain L Q 103 (cited in note 6). But see Ellman, 77 Cal L Rev at 47-48 n 140.Ellman is specifically not concerned with developing a set of external legal or moral normsfor marriage. See id at 49-50. This has also been criticized in the literature. See Schneider,1991 BYU L Rev at 218-19.

278. This is apparent in his assertion that under present rules governing divorce, maritalconduct is distorted in an economic sense, away from the most efficient outcome. Ellman,77 Cal L Rev at 13-14.

279. Ellman argues that alimony should be awarded only upon proof of three elements:(1) that a spouse has made a "marital investment," id at 53-65; (2) that the spouse has

suffered a post-marriage reduction in earning capacity, id at 53-56, 65; and (3) either (a)that the spouse's marital investment resulted in an increase in marital income, id at 65-71,or (b) that the marital investment included primary responsibility for the care of children,id at 71-73. Ellman's approach seems to require that "compensable losses" from marriage

be carefully measured. See id at 49, 78-80. However, he also argues that precise measure-ments are not necessary. See text accompanying notes 290-91.

280. Id at 42, 50 (cited in note 6).281. Id at 47. The paradigm case for which Professor Ellman's formula provides

compensation is one in which the "lower earning spouse," usually the wife, makeseconomic sacrifices to take care of children or to increase the couple's combined income.Typically, these sacrifices might include assuming a larger share of domestic responsibilitiesand yielding career interests when there is a conflict between their jobs. Id at 46-49; seealso Schneider, 1991 BYU L Rev at 246-47 (cited in note 232). June Carbone points outthat under Professor Ellman's formula,

the prototypical award will go to a woman who interrupts a promising career tocare for her children. The woman who fails to develop her earning potential beforethe children are born or her husband's transfer takes effect, the woman withoutchildren who marries a man with an established career, and the man who marriesa higher earning woman will remain financially at risk from divorce.

Carbone, 43 Vand L Rev at 1493 (cited in note 36) (citations omitted).282. See Ellman, 77 Cal L Rev at 19-20, 74-76.

1995]

566 Roundtable

usually presented as the measure of her damages in the event of a breach. LloydCohen describes the marriage contract as a contract for a stream of futureservices, noting that the specifics of these services depend on the social class andcultural norms of the individual husband and wife as well as the exigencies oftheir shared life.283 When a marriage ends in breach or divorce, these streamsof services are lost.284 Cohen states: "[t]he value of this loss is the cost of find-ing a replacement spouse of equivalent value.12 Cohen describes these specificcomponents of the loss: transactions costs of the search for a new partner;stochastic changes in a party's value in the marriage market; changes that resultfrom investment in the prior marriage, such as having children; and the effects oftime on the value of services each spouse has to offer.286

Professor Cohen and other economic theorists recognize that there is a widegap between describing this loss and measuring it.287 Cohen describes the com-putation of damages as a more difficult task than the determination of breach;as he puts it, the court must "determine the future steam of quasi-rents that thenonbreaching party had a right to expect prior to the breach and award thepresent value of that future stream to the nonbreaching party."288 MichaelTrebilcock and Rosemin Keshvani prescribe this formula for compensation of theopportunity costs of a wife's investment in household production:

the present value of the reduction in earnings she has experienced up to thedate of divorce because of her investments in household production (bothbecause she has not worked or has accepted jobs with below opportunity-cost levels of earnings or because she has already suffered a loss in her full-time earnings because of her reduced investment in her own human capital)plus the present value of her future loss of earnings resulting from herreduced investments in human capital minus any gain in consumptionbenefits that she has obtained, up to the date of divorce, from the incre-mental income earned by her husband (because she has freed his time toconcentrate on his career).289

283. Cohen, 16 J Legal Stud at 273 (cited in note 6).284. Id.285. Id.286. Id.287. See, for example, id at 298 ("On a theoretical level, the value of the loss is fairly

straightforward-it is the replacement cost of a spouse of equivalent ex ante value. It isoperationalizing this theoretical measure that is intractable."); Cornell, 26 Fam L Q at134-35 (cited in note 6) (arguing that attempting to evaluate the cumulative effect of theseadaptations on earning capacity may be difficult); Ellman, 77 Cal L Rev at 78-80 ("Thedifficulties involved .. . are significant but not fatal."); Parkman, No-Fault Divorce at140-41 (cited in note 6) ("Wouldn't it be difficult to estimate the effect of marriage onhuman capital?").

288. Cohen, 16 J Legal Stud at 302. He also states that "the damages may be so greatthat the one who breaches the contract is effectively judgment proof." Id at 303. Cohenconcludes, "[o]nly the shortcomings of all other ways of treating divorce make thisalternative at all plausible." Id.

289. Trebilcock and Keshvani, 41 U Toronto L J at 556 (cited in note 6).

[2:S17

Economics and the Problem of Divorce 567

Ira Ellman advocates a different formula, which is similar to the opportunitycost approach but which does not attempt to reduce an award to reflect the levelof consumption during marriage.29 Ellman describes the losses he would com-pensate:

it is a reliance loss, it is the residual loss a spouse is left with after themarriage has ended (without credits or debits from inequities in the ex-change during marriage), it is a financial loss, and it is a loss arising frommarital sharing behavior, which is in turn defined as either child care orbehavior that allows one's spouse to realize a financial gain.29

The problem of imbalances in the parties' exchanges during marriage isespecially difficult to resolve."2 Two other issues have also proved problematic.One is the problem of whether damages should extend to nonfinancial losses,such as a claim arising from declining marriageability.2 93 Another recurring

290. ElIman, 77 Cal L Rev at 54-56 (cited in note 6). Ellman would require that a wifeestablish that her investment led to an increase in marital income, id at 65-71, or that itwas devoted to primary responsibility for child care, id at 71-73. See also Parkman, No-Fault Divorce at 132-34 (cited in note 6) (proposing use of property settlement "adjust-ments" to compensate for reductions in a homemaker's human capital).

291. Ellman, 1991 BYU L Rev at 270-71 (cited in note 82). Kristian Bolin argues that"it is not possible to provide efficiency-promoting rules without information about theagreement concerning the division of the marital income." Bolin, 14 Intl Rev L & Econat 501 (cited in note 5).

292. Compare Trebilcock and Keshvani, 41 U Toronto L J at 556, with Ellman, 77 CalL Rev at 54-56. For an elaboration of Ellman's claim that alimony awards should nottake into account any imbalances in the parties' exchanges during marriage, see Ellman,1991 BYU L Rev at 277-81; Ellman, 77 Cal L Rev at 55-56. This argument is disputed;see Schneider, 1991 BYU L Rev at 211-17 (cited in note 232) (arguing that Ellman'sscheme ignores the possibility that losses were compensated during the marriage). Ellman'sview is that we cannot determine the values of what was exchanged between husband andwife during their relationship, and it is on this ground that he rejects the possibility ofdeveloping restitution-based remedies. ElIman, 77 Cal L Rev at 24-28; but see Carbone,43 Vand L Rev at 1471-85 (cited in note 36) (characterizing Ellman's proposals asrestitutionary). See also Daniel D. Polsby and Martin Zelder, Risk-Adjusted Valuation ofProfessional Degrees in Divorce, 23 J Legal Stud 273, 282-85 (1994) (arguing that thetransaction costs of attempting to measure the values exchanged during marriage wouldapproach infinity). For a fictional account, see Amy Tan, The Joy Luck Club 149-65(Putnam, 1989) (In a story called "Rice Husband," Lena St. Clair describes the systemthat she and her husband Harold have developed for tracking and dividing their expens-es.).

For Ellman, the remedy available to a spouse for unfair exchanges during marriageis to renegotiate or leave the relationship. Ellman, 1991 BYU L Rev at 280-81. This is,of course, only feasible under a legal regime that permits unilateral divorce. Although thefinancial remedies he proposes for long-term losses should make exit a more realisticoption for the lower earning spouse, Ellman recognizes that violence and other types ofcoercion may prevent this course. Id at 281. He suggests that tort remedies are bettersuited to this situation. See id at 298 n 77.

293. See, for example, Cohen, 16 J Legal Stud at 284-87 (cited in note 6). AlthoughEllman is concerned with claims based on lost marriage prospects, Ellman, 77 Cal L Rev

1995]

568 Roundtable

issue in the literature is described in terms of moral hazard: remedies should notbe overly generous, lest the homemaker be rendered indifferent to whether themarriage succeeds or fails.294

These issues play a significant part in determining the correct measure ofdamages for a marital breach. In some accounts, the debate is characterized interms of reliance and expectation measures of recovery.29 s In effect, LloydCohen argues for expectation recovery, following a determination of breach. IraEllman, on the other hand, argues against attempts to resolve the breach ques-tion and therefore for a limitation to financial losses from reliance. MichaelTrebilcock argues that if fault determinations are not a part of the process, theoptimal measure would be one that falls between expectation and reliancerecoveries. He notes that reliance measures suggest a spouse should be left noworse off than if the marriage had never occurred, and expectation measuressuggest that the spouse should share fully in the other's economic returns.296

June Carbone argues that the debate demonstrates a tension between therisks of underperformance and overreliance. Drawing a parallel between theeconomic analyses on divorce and contract remedies, she points out that there isno single "efficient" solution to the problem of contract remedies.297 Similarly,the law of marriage and divorce pulls in two different directions.29 On one

at 80-81 (cited in note 6), he rejects other types of nonfinancial claims, id at 56-58(nonfinancial losses not included in his formula) and 58-65 (only financially rationalsharing behavior included). Compare Cornell, 26 Fam L Q at 130-35 (cited in note 6).

294. See, for example, Trebilcock and Keshvani, 41 U Toronto L J at 557 (cited innote 6). Trebilcock and Keshvani write:

In constructing the hypothetical contract at the time of marriage, would the twoparties agree that the wife should share in the husband's economic returns bothwhere there is no marriage dissolution and where there is? ...[N]o insurer wouldwrite such a policy because the wife may well be rendered largely indifferent to sus-taining or terminating the marriage, given the assumption that her entitlements ondivorce are not contingent on proof of absence of fault on her part for themarriage dissolution. This is the standard problem of moral hazard in insurancemarkets, or in the present context reciprocal opportunism.

Id. See also Peters, 76 Am Econ Rev at 438, 443-44 (cited in note 5).295. Trebilcock, The Limits of Freedom of Contract at 46-47 (cited in note 60).296. Id at 47. With less, the lower-earning spouse will not agree to specialize in

household production; with more, the spouse will no longer be committed to making themarriage succeed.

297. Carbone, 43 Vand L Rev at 1485-90 (cited in note 36); see also Carbone andBrinig, 65 Tulane L Rev at 959 (cited in note 190).

298. Carbone, 43 Vand L Rev at 1487-88. Where the greater policy concern is to deterdivorce, expectation-based remedies will be preferable. Id at 1499-1500; see also note 278.If the goal is to discourage economic reliance on marriage, it is sensible to limit post-marital support to cases of need. This is the effect of most current alimony statutes.Carbone and Brinig describe the "liberal feminist" position as one that stands opposed torules that encourage women's economic dependence on marriage. Carbone and Brinig, 65Tulane L Rev at 992-96. Carbone also argues that where the goal is to encouragedivisions of labor in marriage, a more effective approach is to use sharing principles thatdo not require complex individualized proof, such as income splitting or "partnership"approaches. See id at 1001.

[2:517

Economics and the Problem of Divorce 569

side is the concern with divorce as a social problem, expressed as the concernthat unilateral no-fault divorce allows inefficient breaches of the marital contract,which leads to inefficient arrangements in marriage. On the other, there are theopposite problems that result from economic dependence, and a concern forother values that are also at stake in the process of divorce, expressed in theargument that lifetime alimony awards go too far toward encouraging traditionaldivisions of labor.

This tension highlights a problem that economic analysis tends to repress: itis quite difficult to implement an entirely neutral set of liability rules. Most ofthe proposals offered are not neutral. Explicitly or implicitly, they choose sidesin a policy debate that juxtaposes contradictory concerns.

These types of implicit policy judgments are also apparent in the ongoingdebates over other aspects of divorce remedies. For example, there are economicarguments for property division rules in divorce, designed to encourage optimallevels of human capital investment during a marriage.299 Taking as given thecontemporary rule that "marital property" should be equitably or equallydivided, various writers describe complex methods for valuation of humancapital or "career assets" accumulated during marriage."' The concern here isfor the case in which one spouse leaves the marriage with substantial neweducation or training. Although the remedy is typically described in neutral termsas a reimbursement, the theory behind it incorporates a range of normativejudgments that go well beyond economic efficiency."' These judgments arereflected in the way in which Alan Parkman and others use the human capitalmetaphor when they compare the relationship of husband and wife to a commer-cial loan transaction. 02 When they characterize the supporting spouse's interestin her partner's career or education in terms of debt rather than equity,303 they

299. See, for example, Severin Borenstein and Paul N. Courant, How to Carve a Med-ical Degree: Human Capital Assets in Divorce Settlements, 79 Am Econ Rev 992 (1989);Parkman, No-Fault Divorce (cited in note 6).

300. This analysis builds on a body of judicial decisions, which often characterize theremedy as a restitutionary one. See generally Estin, 71 NC L Rev at 757-67 (cited in note135).

301. As Ira Ellman argues, this is true of all restitution-based theories, which imply ajudgment that there has been unjust enrichment. Necessarily, this implies an understandingof "the social and economic conventions that ordinarily govern the relationship betweenthe parties." Eliman, 77 Cal L Rev at 27 (cited in note 6).

302. The analysis constructs the decision of the nonstudent spouse whether to supportthe other as based on the potential financial returns from this investment, compared to thereturn that would be available on other investments. Parkman proposes that once theamounts of marital contribution are determined, the support amount from each year"should be brought forward to the present at an interest rate which reflects the returnthat the non-student spouse could have recovered if the funds had been invested."Parkman, 40 Ark L Rev at 455 (cited in note 6). See also Borenstein and Courant, 79Am Econ Rev at 1000 (cited in note 299) (proposing a "marginal interest rate rule," inwhich the "investing spouse" reimburses the supporting spouse's support "compounded atthe investing spouse's period-one [educational period] marginal interest rate").

303. Parkman, No-Fault Divorce at 134 (cited in note 6).

1995]

570 Roundtable

are making a significant normative judgment about the meaning of marriage.When they depict the remedial problem as a technical matter of finding thecorrect interest rate, they are missing-or hiding-these issues, and their appar-

ent neutrality hides the implication of their proposals, which is to limit signifi-cantly the scope of the "investor's" recovery.3 °4

The weakness of normative arguments from Kaldor-Hicks efficiency princi-ples is that they require postulating hypothetical consent, which has less compel-ling normative implications than the actual consent in the Pareto model. MichaelTrebilcock notes that it is hard to determine the actual welfare implications ofalternative rules under Kaldor-Hicks norms, and it is not clear that efficiencywould be the basis on which real parties would consent."' Nonetheless, hedefends the economic preference for private ordering of family relationships, inthe belief that this advances both individual autonomy and social welfare."'He rejects both the application of "across-the-board legal norms" for marriageand divorce and the use of "expansive ad hoc judicial discretion" in decidingcases, although he stipulates that private ordering requires an appropriate set ofbackground legal norms.30 7 How can appropriate norms be determined?Trebilcock notes that "economics conventionally takes prior endowments andentitlements as givens," and, from this perspective, it "has limited insights tooffer on the crucial question of fashioning an appropriate set" ofentitlements.3 8 His own answer is revealed in his description of theseentitlements as default rules, 0 9 and the suggestion that they are justified based

304. An equity investor in a career "is entitled to a share of the profits" it generates,whereas a debt lender "is entitled only to a market rate of return on the investment."Parkman, No-Fault Divorce at 134. But see Polsby and Zelder, 23 J Legal Stud at 275(cited in note 292) ("In practice, the O'Brien equity solution and the much more prevalent'debt' solution are unlikely to work out very differently. Both will yield equivalently fairoutcomes and both will be equivalently efficient if the right interest rate is assigned. Butcourts do not seem to have adopted an interest rate that is realistic, on a risk-adjustedbasis, to compensate the [non-degree spouse] adequately for her investment and the riskassociated with it.").

305. Trebilcock, The Limits of Freedom of Contract at 246 (cited in note 60). Nev-ertheless, he posits a "hypothetical contract" based on efficiency criteria in his recommen-dations for divorce entitlements. As Professor Trebilcock notes, the Pareto efficiencyframeworks applied in economic theory shift almost inevitably toward a Kaldor-Hicksapproach in practice. Id. In some situations, the normative basis for exchange is sufficient-ly weak that it is no longer a useful method of regulation. Faced with these difficulties,Trebilcock argues for a legal approach that recognizes a range of values and seeks toidentify which legal processes can best be used to vindicate the different values at stake.Id at 248. He argues that courts are best situated to inquire into the choices and prefer-ences of particular parties, to vindicate autonomy values rather than broad questions ofefficiency or distributive justice. Id at 249. Where more pervasive issues of market failureare concerned-such as monopoly, information failure, externalities, coordination andcollective action problems-Trebilcock advocates regulatory approaches. Id.

306. Trebilcock refers to this as the "convergence claim." Id at 21-22.307. Id at 56.308. Id at 57.309. Trebilcock believes that parties should be free to agree to different terms, em-

[2:517

Economics and the Problem of Divorce 571

on a hypothetical consent. "[W]e may find it useful conceptually to construct ahypothetical contract at the time that marriage is entered into, when the partiesdecide on a specialization of functions involving one partner's devoting time tohousehold production and the other to market production."31 Constructingthe hypothetical contract which Trebilcock postulates clearly requires policyjudgments from beyond the domain of microeconomic theory."1'

C. THE LIMITATIONS OF PROPERTY RULES

Some law and economics analysis recognizes the complexity of the normativeproblems that any efficiency analysis implies. Writing more than two decadesago, Guido Calabresi and A. Douglas Melamed argued:

The first issue which must be faced by any legal system is one we call theproblem of "entitlement." Whenever a state is presented with the conflict-ing interests of two or more people, or two or more groups of people, itmust decide which side to favor. Absent such a decision, access to goods,services, and life itself will be decided on the basis of "might makesright"-whoever is stronger or shrewder will win. Hence, the fundamentalthing that law does is to decide which of the conflicting parties will beentitled to prevail.312

Following Calabresi and Melamed, any meaningful discussion of legal rulesshould begin with a discussion of entitlements.3 13 The question of what formsof rules can best protect those entitlements comes as a secondary matter.

Economists have useful input on both of the central legal questions thatCalabresi and Melamed identify. On the question of what entitlements familylaw should recognize, economic theory would dictate that the entitlements shouldgenerate optimal allocations of resources. But the theory does not indicate whichresources matter. Are we seeking to optimize marital relations, human happiness,

phasizing however the need to "be sensitive to various forms of transaction-specific failuresthat parties may be particularly vulnerable to in times of extreme stress, guilt, remorse,and emotional turmoil." Id at 47.

310. Id at 46.311. Professor Scott's use of the "hypothetical bargain heuristic" assumes that parties

would choose terms that reflect a "commitment to a relationship of reciprocal obligationand sharing" by making divorce more difficult than current laws provide. Scott, 1994Utah L Rev at 723-25 (cited in note 22).

312. Calabresi and Melamed, 85 Harv L Rev at 1090 (cited in note 61). Although thevocabulary Calabresi and Melamed develop is sometimes applied in the economic literatureon divorce, see, for example, Zelder, 16 Harv J L & Pub Pol at 246 n 11 (cited in note5), the subtlety of their insight is not.

313. Calabresi and Melamed discuss the problem of setting entitlements at some length,Calabresi and Melamed, 85 Harv L Rev at 1093, 1105 (cited in note 61), although theirlarger concern extends to the types of rules applied to protect different types ofentitlements. See text accompanying notes 60-69. See also Macneil, 68 Va L Rev at 961-64 (cited in note 35) (discussing legal entitlements as part of the "underlying relationalstructure" which allows deal-making).

1995]

572 Roundtable

or divorce? Do we care about the allocation of fertility, beauty, wealth, loyalty,and intelligence?

The theory also suggests that we should be concerned with reducing transac-tion costs. This is also not a straightforward question, as the relevant "costs"come in many different varieties, and many are not monetary. Further, asCalabresi and Melamed point out, some of these considerations are not readilyreduced to the terms an efficiency analysis requires. What of the effects of agiven allocation or set of costs on distributions of "wealth"? What of the effectsof these allocations of costs on third parties? Which third parties, and whicheffects, should we take into account? Are a paramour's interests to be factoredinto the calculus?

On the question whether rights should be protected with property rules,liability rules, or inalienability rules, economic theory similarly offers only apartial response. Property rules are normally more efficient when one party isbetter situated to balance the costs and benefits of a particular entitlement. If,however, we are uncertain which party is the cheapest cost avoider, liability rulesare better.314 Moreover, property rules appear more efficient whenever transac-tion costs are low enough to permit the parties to alter the initial allocation ofa right by private agreement. Where this is not the case, liability rules are oftenpreferable, because they allow "the economic efficiency of a proposed transfer ofentitlements to be tested" through the mechanism of damages awards. 3 Hereas well, we need a more detailed analysis of what transaction costs exist, and aframework in which they can be measured and compared.

The literature on mutual consent divorce does pose the question whether anentitlement to marriage for life or marriage at will is preferable. The argumentis made that an entitlement to marriage for life is preferable on efficiencygrounds (it leads to an optimum level of divorce);1 6 on distributional grounds(it will shift resources toward those who oppose divorce, who are more likely toneed them);3"7 and on other justice grounds, including protection of thirdparties (i.e., children) and the moralism that disfavors divorce.1 ' What themutual consent argument neglects is the next step: Why should an entitlement tocontinue marriage for life be protected with a property rule?3"9

Where the concern is with how property rights are specified, efficiencyarguments suggest that we should consider the possibility of negative externalities

314. Calabresi and Melamed, 85 Harv L Rev at 1119-20. Ian Ayres and Eric Talleyhave challenged this "conventional wisdom." Ian Ayres and Eric Talley, Solomonic Bar-gaining: Dividing a Legal Entitlement to Facilitate Coasean Trade, 104 Yale L J 1027(1995).

315. Calabresi and Melamed, 85 Harv L Rev at 1127.316. See text accompanying notes 85-87 and 101-05.317. See text accompanying notes 122-24.318. See text accompanying notes 97-98 and 133-37.319. While Becker has not addressed this question, Zelder collapses the two steps

together, asking whether a property right in marriage is preferable to a property right indivorce. See text accompanying notes 113-16. Parkman's argument is that property rulesare preferable in order to internalize the nonfinancial costs of divorce. See note 131.

[2:517

Economics and the Problem of Divorce 573

and a corresponding over-utilization of the rights.32 Economic criticisms of the"property right" in divorce are based on these effects. But a property right inmarriage, which is in effect a right in one's partner's ability to enter a newrelationship, might also be "overutilized," and this issue seems not to be consid-ered.

In the contract enforcement literature, the discussion of entitlements inmarriage is less explicit. There are references to protecting and encouragingmarriage, but the specific entitlements that are debated tend to be more qualified.Not every spouse unhappy with the prospect of divorce can claim protection foropportunity losses or the destruction of quasi-rents. Here, the question of howa limited entitlement could be protected is the more central concern.

Following the Calabresi-Melamed approach, it is very important to discoverdifferences in the costs of various rules and entitlements."' This is a very com-plex project. In his prior writing, Judge Calabresi described different types ofcosts of accidents and pointed out that some rules might reduce one set of costswhile increasing another."z As with accidents, we may wish to reduce theincidence of divorce itself, as a type of primary cost."~ In addition, however,we may be concerned with the secondary or societal costs of divorce, and alsothe tertiary or administrative costs of the process. 24 These costs are of signifi-cant concern in the context of divorce."2 The argument for mutual consentdivorce addresses primarily the first cost: reducing the frequency of divorce.3 2 6

In the case of marriage and divorce, there is no evidence that one party orthe other is in a better position to value the gains and losses from marriage ordivorce. There is also no reason to believe that we can identify, with a generalrule, the party better situated to avoid the losses divorce entails. Because there isno market in which family relationships and personal freedom can be valued orcompared, transactions that require valuation of these interests are likely to bedifficult and expensive, and subject to many forms of market and contractingfailure. There is also no evidence that these transactions are readily negotiat-ed.

3 27

320. Trebilcock, The Limits of Freedom of Contract at 10 (cited in note 60).321. They use illustrations drawn from nuisance law and the law of accidents, as well

as a discussion of criminal law rules. Macneil stresses this point in his analysis of contractremedies. Macneil, 68 Va L Rev at 957 (cited in note 35).

322. Guido Calabresi, The Costs of Accidents: A Legal and Economic Analysis (Yale,1970).

323. Id at 26-31.324. Id. Twila Perry reviews Calabresi's analysis of the types of accidents costs in her

argument for a "strict liability" tort approach to the losses from divorce. Perry, 52 OhioSt L J at 68-70 (cited in note 266).

325. One understanding of the abandonment of fault-based divorce litigation is as ajudgment that the tertiary costs of the system were so great as to outweigh any benefitthey might have in reducing the primary costs or level of divorce.

326. Professor Becker's argument, however, implies that a property rule in marriagewould effectively minimize the other types of costs. This claim is not plausible, however.See text accompanying notes 119-150.

327. Indeed, most of the literature ignores the complex question of costs. See text

1995]

574 Roundtable

Liability rules also involve costs, primarily the cost of collectively valuing andenforcing damages awards for violations of an entitlement.3 2

1 Where these rulesare well elaborated, however, they can facilitate the process of private negotia-tion and exchange.329 In areas of highly subjective preferences, we might see asan advantage that collective valuation "makes it easier to value the costs at whatsociety thinks they should be valued by the victim instead of what the victimwould value them in a free market if such a market were feasible."33 In effect,the collective determinations authorized by liability rules substitute for thecollective valuation processes of the marketplace.

There is no reason to believe that economists or lawyers can determine anefficient set of entitlements and rules without intensive comparisons of differentsystems and their comparative costs and effects. Moreover, unless reforms reducethese costs, changes to the present system will have primarily distributiveeffects.33" ' But there is an important structural difference between property rulesand liability rules that suggests the latter may be much more useful in modulat-ing the complex, non-market entitlements characteristic of family law.332

One important advantage of liability rules is that they serve to define andlimit the scope of particular entitlements. This is illustrated by the debate overwhat liability rules should be available to protect an entitlement in marriage.33

Under any version of these rules, an unhappy spouse is free to seek a divorce.But in the articulation of the details lie significantly different understandings of,and protections for, marriage. The entitlement is more precisely defined by theavailable grounds for divorce, what spousal support payments may be ordered,and how the process is conducted. 34

In the context of strongly competing goods, the greater flexibility of liabilityrules offers a strong advantage. Property admits only one owner, and in the

accompanying notes 406-17.328. Calabresi and Melamed, 85 Harv L Rev at 1119 (cited in note 61).329. Id at 1109.330. Id at 1110 n 40. See also Linzer, 81 Colum L Rev at 119 (cited in note 156)

(discussing the role of specific performance in assuring compensation for nonmarket orsubjective values that are undercompensated by typical damages awards).

331. See Guido Calabresi, The Pointlessness of Pareto: Carrying Coase Further, 100 YaleL J 1211, 1227-28 (1991). And, although distributive concerns are clearly very importantin family law, it is not apparent that efficiency norms have anything useful to offer onthis front.

332. Ayres and Talley argue that liability rules facilitate a greater level of bargainingthan property rules in situations in which the parties have private information about howmuch they value an entitlement. Ayres and Talley, 104 Yale L J at 1058-61 (cited in note314).

333. See text accompanying notes 246-311.334. This is well understood in tort, where both the polluter and the neighbor, or the

driver and the pedestrian, have recognized interests and some means of vindicating theircompeting claims. As Judge Calabresi points out, "[o]ur society is not committed topreserving life at any cost." Rather, we make a judgment about "how far we want to goto save lives and reduce accident costs." Calabresi, The Costs of Accidents at 17-18 (citedin note 322).

[2:517

Economics and the Problem of Divorce 575

context of family life, ownership is a problem. Liability rules seem far bettersuited to balancing contradictory interests.3

- Property-based entitlements neces-sarily chose to protect one interest over another in a pair or series of competingclaims. Once the entitlement is set, property rules accord complete deference andprotection to that interest. This is the most serious drawback of mutual consentproposals: they give no play to the values of personal freedom that have come tooccupy an increasing place in the social understanding of marriage and di-vorce.

336

The tension between these values presents an extremely complex problem incontemporary society. Before the "divorce revolution" that led to no-fault laws,these competing values led to widespread evasions of the terms of divorce laws,creating significant pressure for change in the law. 337 The claim to freedom andautonomy in one's private life comes not from law, but from the culture we nowinhabit.33 Michael Walzer describes the claims of personal freedom in marketterms, contrasting this "sphere of private affairs" with the sphere of kinship,where older norms and constraints apply:

The sphere of private affairs can never be a stable place. The market incommodities works because the men and women who trade in commoditiesare connected elsewhere (most often to their families). But here men andwomen trade themselves, and they are radically disconnected, free-floatingsubjects. It is a way of life that most people will choose, if they have achoice, only for a time.... To say this is not by any means to defendpolitical interventions in private affairs. 'Because we freely love, as in ourwill to love or not,' all such interventions are barred: they represent theexercise of power outside its sphere.339

Walzer's analysis highlights the fact that our ideas of freedom have come toinclude goods once allocated on different principles.34 °

335. The complexity of this task is apparent in the literature on divorce. See textaccompanying notes 243-311.

336. See generally Lawrence M Friedman, The Republic of Choice: Law Authority andCulture 175-78 (Harvard, 1990); Lawrence Stone, Road to Divorce: England 1S30-1987413-16 (Clarendon, 1990) (discussing the "revolution in marital expectations"); Schneider,1994 Utah L Rev 503 (cited in note 22); Carl E. Schneider, Moral Discourse and theTransformation of American Family Law, 83 Mich L Rev 1803 (1983).

337. See Jacob, Silent Revolution at 33-35, 50-51, 66-69 (cited in note 263); Nelson M.Blake, The Road to Reno: A History of Divorce in the United States (Macmillan, 1962);Max Rheinstein, Marriage Stability, Divorce and the Law (Chicago, 1972); Lawrence M.Friedman and Robert V. Percival, Who Sues for Divorce? From Fault through Fiction toFreedom, 5 J Legal Stud 61 (1976).

338. See, for example, Schneider, 1994 Utah L Rev at 522-32.339. Michael Walzer, Spheres of Justice: A Defense of Pluralism and Equity 238-39

(Basic Books, 1983) (quoting Milton's Paradise Lost).340. Id at 237 ("Thus the old laws against copulation, extramarital sex, are understood

[today] as infringements of individual freedom.... [But] they become tyrannical onlywhen physical love is publicly conceived as a good-in-itself. Or, when it is conceived asa good instrument to free choice in marriage. . .).

1995]

576 Roundtable

Rules that make divorce transactions more difficult to achieve wouldpromote marriage at the expense of other, important, social goods. In a propertyrights approach to marriage, the policy choice is formulated as a choice betweencompeting entitlements, in which either "marriage" or "personal freedom" mustprevail, but any explanation of why one set of values need be subordinatedcompletely to another. Certainly, stability in marriage is important in our society,but so is happiness in marriage. Karl Llewellyn argued sixty years ago thatincreasing divorce rates could be viewed as evidence of an improvement in theinstitution of marriage.34' While we may regret the decline of moral discourseand the waning ethic of responsibility,342 there is very little to suggest thatmany of us are willing to resume older attitudes toward intimate relationships.

To the extent that the shift from fault divorce to no-fault divorce representeda move from one property rule to its opposite,343 it is no wonder that bothsystems have seemed unstable. The history of family law rules over the centuriesincludes numerous examples of difficulties encountered by efforts to constrainpersonal autonomy. Informal marriage persisted despite efforts to prohibit it.344

If divorce was impossible in one jurisdiction, it was pursued in another;34 ifdivorce was difficult everywhere, abandonment and bigamy became significantproblems. 3" A move toward property-based rules intended to protect familylife would be at least equally problematic today.347

Liability rules have much greater potential in the setting of marriage anddivorce regulation because they allow for greater attention to the complexnormative dimensions of family life. Economics will more readily achieve itspotential to assist in the design of useful legal rules if this wider framework ismore fully investigated. Such a project also requires a more complete understand-ing of the many different levels at which present legal rules operate, both inregulating the processes of family interaction and in setting backgroundentitlements of family life."

341. Karl N. Llewellyn, Behind the Law of Divorce: II, 33 Colum L Rev 249, 262-63(1933) ("It is urged here that all likelihood tells in favor of the rising divorce rate indicat-ing not decay, but progress in marriage.").

342. See generally Schneider, 83 Mich L Rev 1803 (cited in note 336).343. I have doubts whether fault-based divorce rules are appropriately characterized as

creating a property interest in marriage. See text accompanying notes 97-100.344. Michael Grossberg, Governing the Hearth: Law and the Family in Nineteenth-

Century America 83-102 (UNC, 1985).345. Roderick Phillips, Putting Asunder: A History of Divorce in Western Society 473-

78 (Cambridge, 1988) (describing migratory divorce patterns in the 19th century).346. Stone, Road to Divorce at 141-48 (cited in note 336); Phillips, Putting Asunder at

279-313.347. See text accompanying notes 168-76.348. For illustration of the type of empirical study that would further this goal, see

Brinig and Alexeev, 8 Ohio St J Dispute Resol 279 (cited in note 5).

[2:517

Economics and the Problem of Divorce 577

Ill. The Work of Family Law

Section II of this Article argued that both of the economic approaches to theproblem of divorce require significant normative input from beyond the realm ofefficiency. This has been obscured in most economic analyses of marriage anddivorce, which typically take for granted the larger framework of legal and socialregulation of family life. Economic models would be a more useful basis forpolicy recommendations, however, if these aspects of law's work were more fullyconsidered.

Economic analysis of the family depends on prior specification of a varietyof types of legal rules. The economists' normative preference for private orderingassumes that agreements will be voluntary and fully informed, a standarddifficult to attain in the marketplace and even more elusive in the family.349

Similarly, the economic model recognizes that efficiency goals may not be servedif the private agreements of husband and wife have effects on those "outside" theexchange.35 0 The law plays an important role in setting conditions for consentand defining which external effects should be internalized.

Some family law principles, such as those governing marital property or childsupport, establish the framework of entitlements which shapes the bargains andbehavior of individuals within families. Other rules, such as those governingprenuptial and separation agreements, regulate the processes of family interac-tion. In these respects, the work of law precedes any efficiency analysis ofmarriage and divorce. But the role of law should not be treated as limited tosetting the stage for economic theorizing.

Part of the work of family law is practice, which mediates the generalities oftheory and the particularities of peoples' lives. To the extent that economic workon divorce has considered the applications or specifics of legal rules, it has doneso at a high level of abstraction. Mutual consent and fault-based divorce arelumped together;"5 ' proposals for complex alimony and prop.erty division rulesare developed with little consideration of the effects on divorce practice;1 2 andmost of the literature focuses on two or three paradigm cases, usually withoutevidence to suggest either that these cases are a valid basis for generalization orthat they present more serious problems than other fact patterns.353 If econom-ics is to offer normative analysis, it cannot ignore these subtleties.3 4 This isparticularly important in family law, which is widely viewed as the area of law

349. As Trebilcock argues, both of these conditions imply some baseline moral and legalunderstanding of consent. Trebilcock, The Limits of Freedom of Contract at 242 (cited innote 60).

350. Id.351. See text accompanying note 98.352. See text accompanying notes 283-94.353. See text accompanying notes 133-37 and 186-90.354. This seems particularly important to remember in view of the tendency in the

economic literature to criticize the no-fault divorce reforms for their unintended effects.See, for example, Parkman, No-Fault Divorce at 1-2 (cited in note 6).

1995]

578 Roundtable

least capable of discipline, synthesis, theory, or principle."' In order to gener-ate constructive policy suggestions, economic models must marry the elegant andabstracted realm of theory with the messy realities of practice. Many of theseissues can be described in economic terms as transaction costs or as problems ofcontracting or market failure. But there is another significant aspect of law'swork in the family which is harder to translate into economic language.

Law is also the framework with which we regulate the tension between theeconomic and moral dimensions of family life. This tension is a longstandingproblem in family law, and despite the aspirations behind no-fault divorce, it isnot likely to go away. The tension exists because individuals construct theirprivate relations within a variety of normative structures and with a wide rangeof motivations, not all of which are consistent with the norms of efficiency orrational maximization. The two economic solutions to the problem of divorcerepresent opposing approaches to this problem as well.

A. CONTRACTING FAILURES

In his writing on freedom of contract, Michael Trebilcock emphasizes thatthe private ordering paradigm depends upon law to correct for a variety ofcontracting and market failures. He defines market failures as those problemscharacteristic of "generic classes of exchange activity," and contracting failuresas the problems that arise in specific disputes between two parties.3"6 In thefamily law setting, rules designed to address both contracting and market failuresare common.3"7 There are relatively extensive sets of rules and principles de-

355. Compare Mary Ann Glendon, Fixed Rules and Discretion in Contemporary FamilyLaw and Succession Law, 60 Tulane L Rev 1165, 1168-83 (1986); Carl E. Schneider, TheNext Step: Definition, Generalization, and Theory in American Family Law, 18 J LReform 1039, 1040-48 (1985). This belief has been prevalent in legal writing on marriageand divorce since at least the legal realist period. See Karl N. Llewellyn, Behind the Lawof Divorce: 1, 32 Colum L Rev 1281, 1282-84 (1932).

356. Trebilcock, The Limits of Freedom of Contract at 22 (cited in note 60). Hesuggests that adjudication is best suited to vindicating exchange values in particular casesof market failure, while legislative or regulatory controls are more useful for correctingmarket failures. Id.

357. Problems of contracting and market failures are acknowledged in the economicliterature on marriage and divorce. For discussion of contracting failures, see, for example,id at 47-48 ("[W]e must be particularly sensitive to various forms of transaction-specificfailures that parties may be particularly vulnerable to in times of extreme stress, guilt,remorse and emotional turmoil."); Becker, A Treatise on the Family at 327-31 (cited innote 3) (information and uncertainty problems); Trebilcock and Keshvani, 41 U TorontoL J at 544-45 (cited in note 6) (fraud, duress, and overreaching); Cohen, 16 J Legal Studat 272-73, 291, 298 (cited in note 6) (information and uncertainty). For discussion ofmarket failures, see, for example, Posner, Economic Analysis of the Law at 143-46 (citedin note 2) (effects on children as externality of divorce decision); Trebilcock and Keshvani,41 U Toronto L J at 559 (externalities); Allen, 13 J Econ Beh & Org 171 (cited in note5) (analyzing legal regulation of marriage as device for reducing transaction costs); Beckerand Murphy, 1 J L & Econ at 13-14 (cited in note 6) (effects on children as externalityof divorce); Weiss and Willis, 3 J Labor Econ at 278-79 (cited in note 5) ("public goods"

[2:517

Economics and the Problem of Divorce 579

signed to address problems of consent, such as issues of capacity, duress, andundue influence; and problems of full information, such as fraud, disclosure, andmistake. These rules are particularly apparent in the laws regulating prenuptialand separation agreements s s in the requirements for contracting a valid mar-riage,.59 and in the provisions for annulment or declaration of invalidity.360

Additionally, various other portions of the legal framework of marriage anddivorce compensate for a variety of market failure problems. Most notably, theseprovisions address external effects of family decisions on children and the publicfisc, but they also extend to certain problems of transferability, bilateral monop-oly, and high transaction costs.3 6'

1. Information and uncertainty.

The understanding of consent central to the law of contracts is generallyconsistent with the economic ideal that an exchange should be fully informed.Classically, this is reflected in defenses to contractual obligation based on fraudand mutual mistake.6 Over time, these defenses have become increasinglybroad and have been joined by affirmative obligations of disclosure363 and rulesdirected to changes in circumstances after an agreement is struck s.36 The disclo-sure rules for commercial transactions do not require full information, or eventhat all information in the possession of the parties be shared.365 Among the

or transferability problems concerning children); Cohen, 16 J Legal Stud at 273-98(transaction costs in marriage market); id at 291 (transferability problems with humancapital); Landes, 7 J Legal Stud at 36, 39-40 (cited in note 5) (transferability problems).

358. For limits on enforceability of prenuptial agreements, see, for example, UniformPremarital Agreement Act S 6, 9B ULA 376 (West, 1987); see generally Clark, The Lawof Domestic Relations at 3-5 and 7-10 (cited in note 99). For limits on the enforceabilityof separation agreements, see, for example, Uniform Marriage and Divorce Act 5 208, 9AULA 170 (West 1987); see also Clark, The Law of Domestic Relations at 759-63; Sharp,132 U Pa L Rev at 1424-28 (cited in note 173).

359. In particular, age requirements and waiting periods are designed to assure thatconsent to a marriage is voluntary. See generally Clark, The Law of Domestic Relationsat 36, 88-92.

360. See, for example, Uniform Marriage and Divorce Act S 208(a)(1), 9A ULA 170(providing for declaration of invalidity in cases of incapacity, force or duress, or fraudinvolving "the essentials of marriage"); see also Clark, The Law of Domestic Relations at99-120 (regarding grounds for annulment); Margaret F. Brinig and Michael V. Alexeev,Fraud in Courtship: Annulment and Divorce, 2 Eur J L & Econ 45 (1994).

361. See text accompanying notes 394-444.362. See, for example, Restatement (Second) of Contracts S 152 ("When Mistake of

Both Parties Makes a Contract Voidable"); id S 164 ("When a Misrepresentation Makesa Contract Voidable"); Posner, Economic Analysis of Law at 109-13 (cited in note 2). Inmarriage law, these are grounds for annulment, which is a type of rescission. See Clark,The Law of Domestic Relations at 143-48 (cited in note 99).

363. See, for example, Restatement (Second) of Contracts 5 161 ("When Non-disclosureis Equivalent to an Assertion").

364. See, for example, id S 261 ("Discharge by Supervening Impracticability"); id S 265("Discharge by Supervening Frustration").

365. In contemporary theory, these rules are understood in the context of the risk

19951

580 Roundtable

most problematic aspects of these excuse principles is the determination of whatrange of information, what varieties of risk, are subject to disclosureobligations.366 Economists debate the costs of acquiring information, of pro-tecting against mistakes, or the possibilities of impossibility and frustration. 6 7

Although a somewhat higher standard for information-sharing is applied tonegotiations around marriage and divorce,"' it is more difficult to determinehow the universe of pertinent information should be defined. Some economicanalysis looks especially at problems of asymmetric information, considering suchissues as what each party knows about the likelihood of divorce or the otherparty's opportunities at divorce. 369

Our present legal rules do not require much information sharing beforemarriage; what is legally pertinent in prenuptial agreements is largely financialinformation,370 or, in the law governing annulments, certain key facts about theparties' reproductive capabilities.37' The relatively narrow scope of the informa-tion that must be disclosed between parties to a marriage or divorce reflects thefact that many relevant judgments are extremely difficult for individuals to make,and harder still to second guess with legal principles.372 Marriage is unavoid-ably uncertain. This is emphasized in the positive economic analysis of divorce,which describes divorce as a function of imperfect information at the beginning

allocation or insurance function of contract. See Posner, Economic Analysis of Law at102-09. This approach is reflected in various provisions of the Restatement (Second) ofContracts; for example, in S 154 ("When a Party Bears the Risk of a Mistake").

366. The terminology used in the Second Restatement of Contracts asks whether anevent or fact was a "basic assumption on which the contract was made," Restatement(Second) of Contracts SS 152 (mutual mistake), 153 (unilateral mistake), 161(b) (nondis-closure), 261 (impracticability), 265 (frustration).

367. See Anthony T. Kronman, Mistake, Disclosure, Information and the Law of Con-tracts, 7 J Legal Stud 1, 2-18 (1978).

368. This is sometimes explained in terms of the "confidential relationship" of husbandand wife or a betrothed couple. See Sharp, 132 U Pa L Rev at 1414-24 (cited in note173).

369. Peters, 76 Am Econ Rev at 438 (cited in note 5). Peters writes:As the marriage continues over time, information about the value of alternatives,should divorce occur, become available. Examples of alternatives include the valueof a potential new relationship, (i.e., remarriage after divorce), and the value ofmarket opportunities which might be different at divorce for women who werehomemakers during marriage.

Her theoretical and empirical analysis suggests that the extent of asymmetric informationis one crucial factor in the efficiency of divorce outcomes. Id at 444. Professor Peters'sempirical results suggest that divorce does occur when it is efficient, id at 452-53, but thisresult has been disputed by Douglas Allen. Allen, 82 Am Econ Rev 679 (cited in note 5).

370. See, for example, Uniform Premarital Agreement Act, 5 6(a)(2), 9B ULA 376 (West1987) (requiring fair and reasonable disclosure of "the property or financial obligations"of a party to a premarital agreement).

371. See Clark, The Law of Domestic Relations at 110-15 (cited in note 99); Brinig andAlexeev, 2 Eur J Law & Econ at 54-57 (cited in note 360).

372. See, for example, Kober v Kober, 211 NE2d 817, 264 NYS2d 364 (NY Ct App1965) (granting annulment to wife of Nazi sympathizer).

[2:517

Economics and the Problem of Divorce 581

of a relationship and changing tastes over time.3 37 This uncertainty creates

risks, particularly in view of the hope that the relationship will endure for a life-time. Although some of the risks of marriage have been addressed in economicapproaches, 74 the central problems that remain are which risks matter andhow these risks are best allocated between husbands and wives.

Melvin Eisenberg describes "limits of cognition" as a source of bargainingfailure, pointing out that although rational choice theories assume strongcognitive abilities, there are a variety of important limits on cognition.3 75 Henotes that parties are particularly likely to confront these limits when theyattempt to contract for an intensive, long-term, "thick" relationship such asmarriage .31 In the face of these limits, Eisenberg argues that law should notattempt to hold these relationships together when parties encounter unanticipatedcircumstances, but rather "to allow either party the right to easy exit on fairterms."31 Professor Eisenberg takes this approach to prenuptial agreementsbecause it is impossible to predict at the outset the course the parties' lives willtake and the impact the agreement will have if it comes into play. Eisenbergsupports rules that permit a "second look" at the agreement at the time ofdivorce, even if there is no showing that it was unfair at the time it was execut-ed.

3 78

Another type of uncertainty arises when the legal standards within whichparties negotiate are highly discretionary. This is characteristic of contemporarydivorce laws, and this fact is believed to affect divorce bargaining significant-ly.379 These effects are greatly complicated by the attitudes of the partiestoward risk, including particularly the risks of divorce or custody litigation andthe financial risks of life after divorce.80

An additional problem for marital bargaining concerns the difficulties ofvaluation where the various gains and losses of marriage and divorce are con-cerned. Rather than relying on legal rules defining divorce entitlements and ajudicial process in which these entitlements are translated into financial figures,

373. See text accompanying notes 39-42.374. See text accompanying notes 210-23.375. Melvin A. Eisenberg, The Limits of Cognition and the Limits of Contract, 47 Stan

L Rev 211 (1995). The limits he addresses include bounded rationality and rationalignorance, disposition (e.g., optimism), and certain systematic defects in capability. Id at213-15.

376. Id at 251-54.377. Id at 252-54.378. Id at 254-58.379. See Mnookin and Kornhauser, 88 Yale L J at 969-71 (cited in note 148). See also

Parkman, No-Fault Divorce at 44-45 (cited in note 6).380. In one empirical study, Margaret Brinig and Michael Alexeev examined divorce

bargaining under different types of legal rules in Virginia and Wisconsin and concludedthat a much higher portion of the cases went to trial in Virginia, where rules create muchgreater uncertainty. Brinig and Alexeev, 8 Ohio St J Dispute Resol 279 (cited in note 5).The results of divorce bargaining and litigation in the two jurisdictions also varied; inboth states, wives received custody in three quarters of the cases, id at 283 n 27, butwives in Wisconsin received greater shares of marital property, id at 293 tbl 1.

1995]

582 Roundtable

the mutual consent approach depends upon the parties themselves to make thesedeterminations. This is a complex problem when financial costs and benefits arein question, as the literature on human capital demonstrates.8 1 It is remarkableto expect that most individuals would have sufficient knowledge or informationto put accurate prices on the financial aspects of their relationships, and it iseven more difficult to imagine how valuation of the more intangible aspects ofmarriage would proceed. Because human relationships are not primarily viewedas monetary, individuals are unlikely to value the gains and losses of marriageand divorce in these terms.382 The problem is still worse when it comes todetermining in advance the gains and losses to be expected from divorce.

2. Voluntariness.

Voluntariness issues present another area in which there may be transaction-specific failures of consent. Here as well, classical contract principles provideddefenses. These include lack of capacity and duress, and in the context offiduciary relationships, undue influence.3"3 Such principles have correlates indomestic relations law, forming grounds for annulment or declaration of invalidi-ty of marriages, and a basis for setting aside prenuptial agreements, separationagreements, or divorce decrees.384

More controversial modern expressions of these principles are the defensesbased on "economic duress" and unconscionability. Economic duress argumentsrequest that a party's difficult economic circumstances be taken into account inthe legal determination about the enforceability of an agreement, at least wherethose circumstances result from the other party's misconduct.8 ' Similarly,arguments of unconscionability are sometimes premised on the "unequal bargain-ing power" of the parties to an agreement.386 These defenses are controversial,although they have sometimes been successful in commercial settings. 7

Calabresi and Melamed imply that questions of transaction costs are central-ly significant in the definition and elaboration of entitlements. The work of

381. See text accompanying notes 283-94 and Estin, 36 Wm & Mary L Rev at 1058-61 (cited in note 9); see also text accompanying notes 409-17.

382. See note 109.383. See, for example, Restatement (Second) of Contracts §5 14 ('Infants'), 173 ("When

Abuse of a Fiduciary Relation Makes a Contract Voidable"), 174 ("When Duress byPhysical Compulsion Prevents Formation of a Contract"), 175 ("When Duress by ThreatMakes a Contract Voidable"), 177 ("When Undue Influence Makes a Contract Voidable").See also Posner, Economic Analysis of Law at 113-17 (cited in note 2).

384. See text accompanying notes 358-60 and Clark, The Law of Domestic Relationsat 103-05 (cited in note 99).

385. See, for example, Totem Marine Tug & Barge, Inc. v Alyeska Pipeline Service Co.,584 P2d 15, 22-23 (Alaska 1978).

386. See, for example, Restatement (Second) Contracts S 208 ("Unconscionable Contractor Term").

387. Judge Posner rejects these defenses on economic grounds. Posner, Economic Anal-ysis of Law at 116 (cited in note 2) ("Economic analysis reveals no grounds other thanfraud, incapacity, and duress (the last narrowly defined) for allowing a party to repudiatethe bargain that he made in entering into the contract.").

[2:517

Economics and the Problem of Divorce 583

"transaction cost" economists suggests that understanding these factors is espe-cially important in the setting of complex, long-term relationships such asmarriage.38 ' Transaction cost approaches to marital bargaining indicate that theparties' economic opportunities outside marriage are indeed powerful determi-nants of the substance of their bargains.3 9 This might be seen as a problem ofmarket failure, the result of bilateral monopoly and corresponding opportunitiesfor strategic behavior.390 However the problem is understood, it is an extremelytroublesome one in economic and legal theory. In the literature, the questionsometimes appears as one of causation: is it appropriate to use financial awardsin divorce to remedy problems caused at least in part by the different opportuni-ty structures for women in society?". Or is this an arena in which differentlegal or regulatory approaches are more likely to succeed?392

Transaction-specific failure is a particularly significant issue for a mutualconsent/specific performance/property rights approach to divorce. By denying lawa role in monitoring the process of exchange or enforcing standards of behaviorin marriage, the proposals would mandate a process of bargaining but deny thepossibility or significance of fraud, duress, or coercion. But this is precisely thearea in which courts may best be equipped to intervene. Michael Trebilcockdistinguishes the role of law in correcting for transaction specific failures ofconsent from its role in correcting for market failures. He argues that courts areinstitutionally best adapted to vindicating autonomy values or "the particularchoices and preferences of the parties involved." By contrast, more systemicmarket concerns are better suited to collective, corrective, regulatory action.393

B. MARKET FAILURES

Common law principles are described in economic terms as operating tomaximize economic welfare in society. Judge Posner argues that law often servesto bring a system closer to the results that would be produced if there were aneffective market.39 4 On this theory, understanding the effects of externalities,monopoly, and information problems is centrally important to law's work. Withthe extension of market-based economic theories to non-market behavior, these

388. See Pollak, 23 J Econ Lit at 595-96 (cited in note 6); Allen, 13 J Econ Beh &Org at 199 (cited in note 6).

389. See Lundberg and Pollak, 101 J Pol Econ at 992 (cited in note 145).390. See generally Posner, Economic Analysis of Law at 61-63 (cited in note 2) (bi-

lateral monopoly and strategic behavior problems in settlement negotiations); Mnookin andKornhauser, 88 Yale L J at 972-73 (cited in note 148).

391. See, for example, authority cited in Estin, 71 NC L Rev at 752 n 110 (cited innote 135).

392. An interesting assessment of these issues is suggested in Lundberg and Pollak, 101J Pol Econ at 1002-05. See also Rhode and Minow, Reforming the Questions at 191(cited in note 143) (arguing that a commitment to substantive equality between the sexesrequires attention to public policies as well as the law of divorce).

393. Trebilcock, The Limits of Freedom of Contract at 248-53 (cited in note 60).394. Posner, The Economics of Justice at 4-5 (cited in note 29).

1995]

584 Roundtable

problems take on particular analytic importance. Here, the concern is not simplyhow a given market deviates from the ideal conditions which are the startingpoint for theory, but how far the concept of a market is an appropriate modelto utilize. When economists speak of "markets" in marriage and divorce, theterm is understood to be metaphoric.39 Still, in applying the metaphor, it isimportant to recognize that explicit markets in marriage and divorce would berife with many types of market failures. In order for exchanges in such marketsto have normative justification on efficiency grounds, these failures would requirecorrection.396 Understanding the ways in which family behavior fits and doesnot fit the market metaphor becomes particularly important when economictheory is used as a basis for policy pronouncements.

1. Entitlements.

The allocation of entitlements is the foundation for contract. It establishesthe status quo which market transactions are designed to shift.397 But this is avery complex process, and the specifics of these entitlements are closely tied intothe process of exchange. As Macneil argues, even with simple contracts, theprocess of shifting an entitlement from one party to another is not instantaneous.He writes: "during that period of change, which in real life is complex both inbehavior and law, legal entitlements are themselves changing in complexways-various benefits and burdens of ownership and obligation are shifting, butnot all at once."

398

Michael Trebilcock argues that the absence of well-defined entitlements maybe a significant source of market failure in the setting of marriage and divorce,arguing that "a private ordering regime must necessarily operate" in the shadowof an appropriate set of background legal entitlements. 399 From this perspec-tive, once divorce law has articulated the norms governing marriage, the partiescan fairly bargain around them.

Legal rules affect bargaining in a variety of ways. The statutory and case lawgoverning remedies in divorce state a baseline from which parties can start inconstructing their marital or separation agreements."' In addition, these rules

395. See, for example, Posner, Economic Analysis of Law at 142 (cited in note 2)(contending the marriage market is an apt metaphor for the search process prior tomarriage). See Donald N. McCloskey, The Rhetoric of Economics 74-78 (Wisconsin,1985). For uses of the metaphor, see Becker, A Treatise on the Family at 108-29 (citedin note 3); Cohen, 16 J Legal Stud at 278-87 (cited in note 6).

396. Trebilcock and Keshvani, 41 U Toronto L J at 534-35 (cited in note 6).397. At any given point, the status quo is neither efficient nor inefficient; as Macneil

writes, "[e]fficiency and inefficiency apply only to how the parties got where they are, rel-ative to other ways they might have gotten somewhere else." Macneil, 68 Va L Rev at963 (cited in note 35).

398. Id at 967. Allen Parkman seems to recognize this problem when he suggests thatunilateral no-fault divorce could be permitted during the first year of marriage or until thewife becomes pregnant, whichever comes first." Parkman, No-Fault Divorce at 139-40(cited in note 6).

399. Trebilcock, The Limits of Freedom of Contract at 57 (cited in note 60).400. In negotiation jargon, they offer "objective criteria" to rely upon in reaching

[2:517

Economics and the Problem of Divorce 585

are default terms, which can be applied by courts if parties fail to reach agree-ment on particular issues.40' They may also serve an expressive or educationalfunction, informing individuals about important aspects of family relationships,and a channeling function, encouraging certain socially preferable patterns ofbehavior.40 ' The particular allocations reflected in legal entitlements may alsocontribute to the preferences that individuals bring to the bargaining process. 403

These factors make the determination of the substantive content ofentitlements extremely problematic. The law has no power to determine thedistributions of love,4°4 and efficiency norms are of small use on the fundamen-tal questions of moral and financial entitlement in marriage. 4

1 Once the ques-tions of entitlement before, during, and after marriage and divorce are addressed,economic theory may be of substantially greater use in determining how theseentitlements may best be shifted between the parties.

2. Transaction costs.

Among the most significant sources of potential "market" failure in thenegotiations surrounding marriage and divorce is the issue of transaction costs.Many of these costs are relatively direct: bargaining has a cost, the enforcementof bargains has a cost, and policing bargains to prevent contracting failure hasa cost. These might be readily measured, although there seems to be very littleempirical data on the subject.4 6 In addition, there are less tangible costs fromthe process, including emotional injury, and loss of self esteem and communityreputation. It is sometimes argued that changing social mores and the no-faultreforms have reduced these costs for individuals, but this is debated.40 7 Here

solutions, and help define each party's "best alternative to a negotiated agreement"("BATNA"). See Roger Fisher and William Ury, Getting to Yes 84-98, 101-11 (Penguin,1981).

401. See text accompanying notes 195-97.402. See generally Carl E. Schneider, The Cbannelling Function in Family Law, 20

Hofstra L Rev 495 (1992).403. Theorists including Daniel Kahneman and Amos Tversky have demonstrated these

effects. See Eisenberg, 47 Stan L Rev at 218-24 (cited in note 375).404. This is Walzer's central point about the sphere 'of kinship. Walzer, Spheres of

Justice at 228-29, 231-32 (cited in note 339).405. See text accompanying notes 61-66, 227, and 308-11.406. Occasional newspaper articles address the financial costs of divorce proceedings.

See, for example, Jeffrey Leib, Specialists, Tough Economy Driving up Costs of Divorce,

Denver Post 1-E (Aug 1, 1988). There are also many judicial decisions on the question offee awards, and the most dramatic of these cases reach the mass media as well. See, forexample, Ruth Marcus, "War of the Roses" in Supreme Court? Two Californians Fight$3 Million Legal Bill for Marathon Divorce Case, Wash Post A13 (Dec 16, 1991)(describing the fee dispute in the Stanley and Dorothy Diller divorce case, where feesexceeded three million dollars). Lawyers' fees are, of course, only a portion of the directfinancial cost of divorce.

407. See generally Judith S. Wallerstein and Sandra Blakeslee, Second Chances: Men,

Women, and Children a Decade after Divorce 7 (Ticknor & Fields, 1989) ("No-fault di-vorce is a legal concept that has gained acceptance in this country, but I have yet to meetone man, woman or child who emotionally accepts no-fault divorce. In their hearts, people

1995]

586 Roundtable

the evidence is largely anecdotal.408

The details of how a system of financial awards for lost earning capacitycould be administered has not appeared to concern these theorists greatly. To theextent this issue is acknowledged, it is generally dismissed with the observationthat the practical difficulties are no greater than many others already managedin different legal contexts, particularly damages awards in tort.4"9 Though thismay be true, there are many reasons to be dissatisfied with this answer. We havelittle experience with requiring proof of losses from a specialization of laborduring marriage,"' and the data such computations would require is not welldeveloped.4" It is certainly not clear that the legal system addresses these issueswell in the context of tort claims for homemakers who are injured or killed.41

Most troublesome, if this is to become the only or primary basis for financialawards in divorce, it would add greatly to the expense and complexity of mostdivorce cases.

Divorce proceedings are structured very differently than tort litigation. Largeproportions of divorcing individuals appear without counsel,413 in part becausefunding devices such as contingent fee agreements are neither permitted norfeasible in this setting.414 The costs of this litigation, including expert witnesscosts, are paid by individuals (largely in after-tax dollars)4 ' who can rarely

believe in fault and in the loss associated with the decision to end a marriage.").408. See generally Mnookin and Kornhauser, 88 Yale L J at 971-72 (cited in note 148).409. See, for example, Parkman, No-Fault Divorce at 141 (cited in note 6) ("The

techniques for estimating the effect of marriage on the parties' human capital are similarto those used by economists and financial analysts."); Ellman, 77 Cal L Rev at 78-80(cited in note 6) ("[R]ules of law often call for speculative measurements" and "the useof statistical data to approximate probable results . . . is commonly accepted in otherfields of law.").

Although Professor Ellman does not minimize these issues, he seems to conclude thatthey are manageable. He notes:

There is no way to develop precise measures of the theoretically relevant criteria.But while theoretically defensible principles of alimony cannot be translated into self-executing adjudicative rules, they can give judges more guidance than they now havein following a coherent approach.

Id at 53.410. See Estin, 71 NC L Rev at 750-54 (cited in note 135) (discussing Oregon and

New York cases on this issue).411. See Estin, 36 Wm & Mary L Rev at 1030-32 (cited in note 9).412. See id at 1023-35 (concerning tort claims for injury or death of a homemaker).413. See, for example, Robert B. Yegge, Divorce Litigants without Lawyers: This Crisis

for Bench and Bar Needs Answers Now, Judges J 8 (Spring, 1994) ("One of the majorbarriers to pro se litigants is the complexity of the law and its procedures."); ABA Stand-ing Committee on the Delivery of Legal Services, Responding to the Needs of the Self-Represented Divorce Litigant (Jan 1994).

414. ABA Model Rules of Professional Conduct Rule 1.5(d)(1) ("A lawyer shall notenter into an arrangement for, charge, or collect: (1) any fee in a domestic relationsmatter, the payment or amount of which is contingent upon the securing of a divorce orupon the amount of alimony or support, or property settlement in lieu thereof.").

415. In general, no income tax deductions are allowed for personal, living, or familyexpenses, which are defined to include most aspects of attorneys' fees and costs paid in

[2:517

Economics and the Problem of Divorce 587

afford these expenses without significant sacrifice at a time when their financialresources are already strained. Every additional issue which requires the use ofexpert testimony imposes a serious burden on the parties and a correspondingrisk that the analysis offered will simply be a pro forma, "quick and dirty,"highly standardized product. Adjudication of breach and damages issues alsoburdens judicial resources, because of the need for particularized findings of factand an opportunity for appellate review. Writers who recognize these problemsendorse approaches which are far less dependent on precise quantification. 46

And, significantly, the original economic analysis of alimony by Elisabeth Landeswas not an argument for precise measurement of a homemaker's losses; her pointwas that the traditional criteria for alimony awards served as a reasonable proxyfor a homemaker's opportunity losses. 17

3. Transferability.

The use of market language tends to mask the most important source ofmarket failure, which is that this "market" is purely metaphoric. There is nocompetition in the market for divorce, and no currency for measuring and trans-

connection with a divorce, separation of support decree. Treas Reg § 1.262-1(b)(7) (asamended in 1972). There is an exception to this rule that permits a deduction for theportion of a spouse's fees and costs "which are properly attributable to the production orcollection of amounts includible in gross income [as alimony]." Id. See generally Ruth K.Wild, 42 TC 706 (1964) (42 TCM No 51), acquiesced to in 1967-2 CB 4. The problemof allocating legal expenses among different aspects of a representation is less pronouncedin tort litigation, where expenses are rarely paid out of pocket and recoveries for personalinjuries are largely not taxed. See Internal Revenue Code § 104(a)(2) and Treas Reg S1.104-1(c) (as amended in 1970).

416. These range from variations on discretionary need-based determinations to simpli-fied guideline approaches. For the former category, see, for example, Brinig and Carbone,62 Tulane L Rev at 901-04 (cited in note 190); Carbone, 43 Vand L Rev at 1497-1500(cited in note 36); Robert J. Levy, A Reminiscence about the Uniform Marriage andDivorce Act-and Some Reflections about Its Critics and Its Politics, 1991 BYU L Rev43, 70 ("On balance, it seems to me safer to trust judges to make the simple decisionsrequired by a need-based, post-divorce support for dependent nondegreed spouses standardthan the decisions required by a more complex, more indeterminate 'opportunity loss'standard."). For simplified guideline approaches, see, for example, Singer, 1992 Wis L Rev

at 1117-21 (cited in note 16); Sugarman, Dividing Financial Interests at 159-60 (cited innote 51) ("merger over time" approach); Robert G. Williams, Should There Be Child

Support and Alimony Guidelines?, in Alimony 161 (cited in note 251). See also Cohen,16 J Legal Stud at 298 (cited in note 6) ("At the very least, the search for some fixedand automatic rules . . . may well be preferable to the complex regimes that are in voguetoday.").

417. Landes, 7 J Legal Stud at 50 (cited in note 5). Marsha Garrison's empiricalresearch in New York indicates that most present alimony awards can be explained bythree factors: the length of the marriage, the percentage of family income earned by thewife, and whether she had been unemployed for more than ten years. Marsha Garrison,Presentation at International Society of Family Law North America Conference on FamilyRestructuring at the End of the Twentieth Century (June 1993). Garrison suggests that abright line test based on these criteria would capture the outcomes of most presentlitigation and provide more guidance for settlement negotiations. Id.

1995]

588 Roundtable

ferring values between the parties. The absence of a common currency meansthat exchanges are very difficult to transact, particularly on those questions thatare subject to only "yes or no" choices. Marital status, for example, is an all-or-nothing proposition. In their important analysis of divorce bargaining, Mnookinand Kornhauser point out the difficulty of bargaining on any issue for whichthere is no middle ground."8 They suggest that the move toward a wider vari-ety of custody options and a greater level of judicial discretion has the effect offacilitating negotiations." 9 But many aspects of family relationships are funda-mentally nontransferable, posing significant obstacles for an exchange-basedtheory of family life.42°

This is a problem that has received serious attention from academics andjudges, who have asked how far parties are likely to "trade" money for greatershares of custody. Although there is disagreement as to how prevalent thepractice has become, the commentators seem to agree that such exchanges aremorally problematic.4 21 The possibility of custody trades also implies a moreserious set of market failure issues for marriage and divorce bargaining: theproblem of the external effects of decisions made by husbands and wives.422

4. External effects.

The social costs and benefits of family life are of particular concern in familylaw. Two varieties of externalities are commonly identified: the positive andnegative effects of family behavior on the larger society; and the effects of familydecisions on individual third parties, particularly children. These effects appearwith some regularity in legal analysis. Public policy restrictions on familybargaining are justified based on the public interest in marriage, put vividly byone court with the declaration that the state is a party to every marriage con-tract.423 The risk that a spouse or child will become a public charge is recitedas a basis for support orders, or restrictions on contracts concerning support

418. Mnookin and Kornhauser, 88 Yale L J at 975-76 (cited in note 148).419. Id at 977-80.420. I suspect that this is one reason for the greater interest among academics in

economic models of divorce. Martin Zelder, in the proposal discussed in the text accompa-nying notes 101-16, argues that bargaining over terms for divorce is more efficient thanarguing over terms on which to remain married, largely because there are many "publicgoods" in marriage. Marital bargaining is evidently much more complex and less amenableto mathematical modeling. See Pollak, 23 J Econ Lit at 603-05 (cited in note 5). Butdeciding on this basis to focus on divorce seems a great deal like the old joke about thefellow looking under a streetlight for the keys he lost somewhere down the block.

421. Eleanor E. Maccoby and Robert H. Mnookin, Dividing the Child: Social and LegalDilemmas of Custody 44-46, 154-59 (Harvard, 1992); Neely, The Divorce Decision at114-18 (cited in note 264); Brinig and Alexeev, 8 Ohio St J Dispute Resol at 287-90(cited in note 5).

422. Judge Posner describes the task of correcting externalities as "the most dramaticeconomic function of the common law." Posner, Economic Analysis of Law at 254 (citedin note 2).

423. Fricke v Fricke, 257 Wis 124, 126, 42 NW2d 500, 501 (Wis 1950) ("There arethree parties to a marriage contract-the husband, the wife and the state.").

[2:517

Economics and the Problem of Divorce 589

obligations, or tougher child support enforcement laws.424 The scope of thisproblem is much greater than legal materials have recognized. 4' Ideally, familylife is a source of positive externalities for the wider society, through the invest-ment in human capital of children and adults,426 and a range of other beneficialeffects.4"7 In the contexts of family dissolution or dysfunction, however, thenegative externalities may be enormous, posing impossibly difficult policy prob-lems. The list includes the distressingly obvious-domestic violence, abuse andneglect of children and the elderly, the child support crisis-as well as a long listof more subtle and disputed social effects of our conduct of family life.

These issues have enormous policy implications. It is particularly difficult toassess how far family law should seek to impose community moral standards onfamily behavior, or to protect individuals against significant self-destructivebehavior. As Michael Trebilcock points out, it is difficult to reconcile the ideathat law should control this type of externality with ideals of individual autono-my and family privacy.42

Even between two parties to a marriage, it is difficult to determine whichtypes of costs should be borne separately by husband and wife and which areappropriately internalized to their marriage. Some economic arguments charac-terize alimony payments as a system for bringing into the divorce decision thecosts of divorce that might otherwise be externalized.4 29

424. This policy is reflected in rules denying enforcement of certain waivers of spousalsupport in premarital agreements. See, for example, Newman v Newman, 653 P2d 728,

735 (Colo 1982) (Health or employability may have so deteriorated in the course ofmarriage that enforcing prenuptial maintenance provisions might "result in the spousebecoming a public charge."); Uniform Premarital Agreement Act § 6(b), 9B ULA 376(West 1983) ("If a provision of a premarital agreement modifies or eliminates spousalsupport and that modification or elimination causes one party to the agreement to beeligible for support under a program of public assistance at the time of separation ormarital dissolution, a court, notwithstanding the terms of the agreement, may require theother party to provide support to the extent necessary to avoid that eligibility.").

425. For descriptions in the economic literature of different externalities present infamily interactions, see sources cited in note 357.

426. Stout, 81 Georgetown L J at 1949-55 (cited in note 207).427. These effects are routinely invoked by the Supreme Court in its opinions on family

issues. See, for example, Boddie v Connecticut, 401 US 371, 376 (1971) ("[M]arriage in-volves interests of basic importance in our society."); Maynard v Hill, 125 US 190, 205(1888) (Marriage has "more to do with the morals and civilization of a people than anyother institution.").

428. Trebilcock, The Limits of Freedom of Contract at 58-69 (cited in note 60).Another perplexing problem is that the range of individual third parties who claim aninterest in particular family decisions has also begun to grow, extending now to grandpar-ents, stepparents, and a bewildering array of biological parents. So far, none of the modelsof efficient divorce has incorporated the interests of these other parties. In the paradigmcase, why is there no consideration of the utility function of a spouse's lover, who cannotmarry that person until a divorce is secured?

429. See, for example, Parkman, No-Fault Divorce at 122-23 (cited in note 6). AsTrebilcock's argument implies, the fact of such externalities is a problem of market failurebest addressed through carefully defined legal entitlements. In contrast, mutual consent

1995]

590 Roundtable

A more significant aspect of this problem concerns the external effects ofmarriage and divorce on children. Bargaining models of marriage and divorcerecognize that there is a problem here, but have not addressed it well.43 Oneapproach simply assumes that the parents-or at least one parent-will behavesufficiently altruistically toward the parties' children to incorporate their interestsinto the agreement.43" ' This may often be more or less accurate, but it is certain-ly not always the case.432 A different model imagines children's support as asubject of negotiations between children and their parents.433 Another, whichis implicit in some accounts of custody and support negotiations, treats parent-child relations as a commodity bargained over between husband and wife.Present legal rules require that those aspects of marriage and divorce bargainingdirectly affecting children be resolved in the best interests of the child, but thisstandard is notoriously difficult to pin down.434

Proponents of both the mutual consent and enforcement proposals assertthat their reforms would benefit children by making it financially safer forwomen to compromise career goals in order to devote their energies to their chil-dren. 35 Implicit in this claim is a view that current divorce practices havesignificant negative external effects on children. For the most part, however, thecorrelation between women's interests and children's needs is asserted ratherthan analyzed. There is some interesting empirical data to suggest that despiterelatively open-ended divorce statutes, families continue to adhere to a norm ofmother-custody and father-support.4 36 But the literature on custody indicates

proposals with no framework of rules to define which subjects are appropriate forbargaining would internalize these costs indiscriminately.

430. References in the economic literature to effects of divorce on children includePosner, Economic Analysis of Law at 143-45 (cited in note 2), and Becker and Murphy,31 J L & Econ at 12-14 (cited in note 6).

431. See, for example, Landes, 7 J Legal Stud at 36 n 3 (cited in note 5).432. One interesting study explored the motivations of parents of preschool age children

who agreed to joint custody arrangements. In some cases, these arrangements reflected theparents' lack of a strong commitment to their child's well-being. Rosemary McKinnon andJudith S. Wallerstein, Joint Custody and the Preschool Child, in Jay Folberg, ed, JointCustody and Shared Parenting 153 (Guilford, 1991).

433. This literature emphasizes the contracting and market failure problems these"transactions" imply; see Stout, 81 Georgetown L J at 1947-48 (cited in note 207); Beckerand Murphy, 31 J L & Econ at 5-8 (cited in note 6).

434. See Carl E. Schneider, Discretion, Rules and Law: Child Custody and the UMDA'sBest-Interest Standard, 89 Mich L Rev 2215, 2219-26 (1991); Jon Elster, SolomonicJudgments: Against the Best Interests of the Child, 54 U Chi L Rev 1, 12-16 (1987);David L. Chambers, Rethinking the Substantive Rules for Custody Disputes in Divorce, 83Mich L Rev 477, 480-86 (1984).

435. See text accompanying notes 121-24 and 205-29. While I support the goal of mak-ing caregiving less financially risky, see Estin, 71 NC L Rev at 746-48 (cited in note 135),I think the economic analysis tends to reduce this problem too quickly to a question oftraditional gender roles.

436. See, for example, Maccoby and Mnookin, Dividing the Child at 112-14, 130-31(cited in note 421); Brinig and Alexeev, 8 Ohio St J Dispute Resol at 290-92 (cited innote 5).

[2:517

Economics and the Problem of Divorce 591

that assuming a correlation between traditional gender roles and the demands ofchildrearing is enormously problematic.437

In the context of divorce bargaining, some writers advocate principles thatwould "put children first," either by limiting access to divorce for parents ofminor children, or by giving children a first call on their parents' financialresources.43 A growing number of cases consider whether children should beallowed to intervene in their parents' divorce proceedings or to retain indepen-dent counsel to represent their interests.43 9 It is far from clear, however, howfar the interests of children should be allowed to override the interests of fathersand mothers. "Children-first" policy recommendations suggest a determinationthat external effects are so significant that bargaining approaches to divorce arenot appropriate.

The issues of contracting and market failure are significant to both visionsof efficiency in marriage and divorce. With a mutual consent divorce rule, basedon a property right in the continuation of marriage, market and contractingfailures may prevent transactions entirely or dramatically shift resource alloca-tions away from an efficient outcome." Under an enforcement and damagesmodel of divorce, an understanding of the effects of these types of problemsseems essential to structuring incentives properly. Yet, although the economic lit-erature on marriage and divorce makes reference to various types of contractingor market failures, they have not been systematically described or discussed."'My argument here is that without a better understanding of these effects, botheconomic "solutions" to the problem of divorce remain highly theoretical, withlimited practical significance.

Michael Trebilcock states the case for incorporating these issues even morestrongly. Moreover, he argues that economic analysis may be particularlyimportant in directing our inquiry to the varieties of factors that prevent pri-vately bargained outcomes from satisfying our normative criteria." He arguesthat the chief areas of moral concern in contract law can be explained in termsof contracting and market failures."3 Thus, economics may be most useful tofamily law to the extent it develops vocabulary and analysis that can be used toexplore the connections between economic and moral issues. To do this, howev-

437. See, for example, Chambers, 83 Mich L Rev 477, 515-24 (cited in note 434);Mary Becker, Maternal Feelings: Myth, Taboo, and Child Custody, 1 Rev L & Women'sStud 133, 203-23 (1992); Martha A. Fineman, The Neutered Mother, 46 U Miami L Rev653 (1992).

438. See Mary Ann Glendon, Abortion and Divorce in Western Law 94-101 (Harvard,1987).

439. See, for example, J. A. R. v Superior Court, 20 Fain L Rep 1417 (Ariz Ct AppJun 28, 1994) (holding seven-year-old entitled to attorney of his choice but not tointervention as a party in contested custody proceeding), and cases cited therein.

440. This is closely related to the more general criticisms of the property rightsapproach, discussed in the text accompanying notes 314-48.

441. But see note 357.442. Trebilcock, The Limits of Freedom of Contract at 29 (cited in note 60).443. Id.

1995]

592 Roundtable

er, will require elaboration of not only the parallels between family and marketrelationships, but also their divergences.444

C. ECONOMIC AND MORAL CLAIMS

Given the range of issues that have not been fully addressed in the economicanalysis of marriage and divorce, it seems at least premature to conclude thatcertain legal frameworks will increase the efficiency of family life. This is aquestion completely separate from the larger question whether, or to whatextent, efficiency norms should set the agenda for family policy. One answer tothis larger question may be that efficiency norms are more useful in resolvingsome particular problems, which we might characterize as those involvingeconomic rather than moral claims.

In the present debates over the rules for divorce, economic and moralarguments are sometimes seen in competition, and sometimes collapsed together.In his writing on divorce, Michael Trebilcock only advocates one specific legalentitlement, which is designed to adjust the human capital consequences ofspecialization in the household.44 Despite his interest in contracting and mar-ket failures, he does not discuss the broader problems associated with householdor divorce decisions. Professor Ellman's use of economic principles is intended tokeep the law neutral on questions of family life. He argues that it is not possible"to devise a functional law of alimony that takes proper account of the nonratio-

nal and nonfinancial goals and motivations involved in marriage, while avoidingdeterminations of fault."446 On the other hand, Professor Schneider criticizesEllman's neutral stance, suggesting that the force of Ellman's paradigm casecomes from its moral appeal.447 Similar criticisms are made by Arthur Cornell,who objects to Ellman's proposal to limit compensation to losses arising from"economically rational" choices.448

In contrast to scholars such as Trebilcock and Ellman, the advocates ofmutual consent divorce are interested in addressing a wide variety of non-financial losses from divorce. As described in the economic literature, these in-clude the loss in marriageability of one spouse,449 the many effects of divorceon children,4"' losses from misconduct during the marriage such as violent

444. Gary Becker has suggested concerns along these lines as well. See Becker, ATreatise on the Family at 303-04 (cited in note 3); Becker, 101 J Pol Econ at 400 (citedin note 1).

445. Trebilcock, The Limits of Freedom of Contract at 46-47.446. Ellman, 1991 BYU L Rev at 270 (cited in note 82).447. Schneider, 1991 BYU L Rev at 247 (cited in note 232).448. Cornell, 26 Fam L Q at 133 (cited in note 6).449. See, for example, Parkman, No-Fault Divorce at 135-36 (cited in note 6); Becker,

A Treatise on the Family at 330-31 (cited in note 3); Ellman, 1991 BYU L Rev at 282-85(cited in note 82); Cohen, 16 J Legal Stud at 278-87 (cited in note 6); Ellman, 77 Cal LRev at 43-44, 80-81 (cited in note 6); Brinig and Carbone, 62 Tulane L Rev at 873-76,894-95 (cited in note 163); Landes, 7 J Legal Stud at 51 (cited in note 5). See alsoParkman, No-Fault Divorce at 7-8 (discussing search costs for a new mate).

450. See Carbone, 43 Vand L Rev at 1488-91 (cited in note 36) (The externalities of

[2:517

Economics and the Problem of Divorce 593

behavior4s and one partner's loss due to "the desire for a continuing relation-ship with the other.452 Parkman does not describe these as moral concerns, buthe argues repeatedly that his goal for mutual consent divorce is to force thespouses to recognize all of these costs when considering divorce.45 3

Why do scholars such as Trebilcock and Ellman limit their recommendationsto the more financial aspects of marriage and divorce? Ellman recognizes thenonfinancial losses of divorce and suggests a range of different reasons foromitting these from his theory. He starts with the fact that "measuring nonfinan-cial gains and losses may be conceptually as well as practically impossible." 4

In addition, he suggests that incentive structures may be less important wherenonfinancial matters are concerned. 5 Professor Ellman also argues that somelosses from divorce are so different in character as to require a finding of faultbefore compensation would be appropriate.4"' He writes: "[a]ny new theoryone might develop to allow recovery for the emotional pain one spouse inflictsupon the other would surely require a finding that the claimant's emotional lossflowed from some conduct of the defendant's that one is prepared to label asblameworthy."4"7 In other words, these are moral claims which are not man-ageable within the parameters of an economic or legal system that attempts to beneutral on the conduct of marriage.

This difference between the enforcement and mutual consent solutions is themost significant. With the creation of a property right in marital status, themutual consent approach allows the parties to incorporate "moral" consid-

what appear to be "economically efficient" marital decisions may include toll on childrenand society, perpetuation of gender discrimination, and restrictions on women's partici-pation in work force.).

451. Ellman suggests that these losses might be better treated with a tort approach.Ellman, 1991 BYU L Rev at 305-06.

452. Parkman, 8 BYU J Pub L at 104 (cited in note 5).453. Parkman, No-Fault Divorce at 8, 104, 123, 138.454. Ellman, 1991 BYU L Rev at 282. He writes:

The problem is not just that measuring the joys of parenthood or the pain ofestrangement are difficult, or that different people place different values on closebonds with their children, even though both observations are true. It is also thatsuch a comparison requires a common scale upon which to measure the financialoutcomes, the burdens and benefits the wife derives from the children, and the painof estrangement suffered by the husband. Does the wife derive more pleasure fromher children than the husband derives pain from his estrangement from them? Suchquestions may not even make sense, regardless of whether we ask them for thepurpose of reallocating losses to insure a proper incentive structure or simply toachieve some kind of equity.

Id.455. Id at 282-83.456. Id at 283-84. Ellman's explanation for omitting the remarriage problem from his

theory is that it "requires a different theoretical exercise" than the one he advances.Ellman, 77 Cal L Rev at 81 (cited in note 6). "Although some remedy is probably neces-sary, we may conclude that the obligation is society's and not the former husband's." Id."Beyond that, there are difficult questions in establishing the amount of such claims, andin determining which wives should be entitled to them." Id.

457. Ellman, 1991 BYU L Rev at 283-84 (cited in note 82).

1995]

594 Roundtable

erations into their divorce decisions. In effect, the entitlement is defined toinclude any aspect of married life that a person feels strongly enough about touse as a basis for withholding consent to divorce. This permits a spouse todemand payment for loss of social status or self-esteem, or loss of a full relation-ship with children, or emotional pain. And, because the determination of whichissues to bargain over is left to each individual, it leaves the state in a position ofapparent neutrality with respect to the moral aspects of marriage. Thus, with itsemphasis on private ordering, the mutual consent regime appears to maintain amorally neutral stance as well. But mutual consent proposals are not morallyneutral. Their purpose is to reduce the incidence of divorce and change itsperceived distributive effects."' Mutual consent proposals in the literaturerepresent a normative judgment that mothers and children are deserving of agreater share of family financial resources on divorce. They imply that divorceshould be a great deal more costly than it is under present laws, and there is theobvious conclusion that if divorce is made more expensive it will become lesscommon. But there is no obvious correlation between this goal and a propertyright in marriage; carefully drawn financial entitlements in divorce could achievethe same purposes more directly.

An advocate might argue that mutual consent rules reflect a moral obligationto keep one's promises. Although this is not the norm in commercial settings,4 s9

the special nature of marital bargains may seem to call for a different rule. Thequestion of moral obligation is an important one, but it is not a normativeeconomic one."' Moreover, a moral rule which admits of no qualification orexception would be seriously flawed.46' What are the arguments for an indirectapproach to the distributive problems of divorce? Writers who consider specificperformance remedies in business settings emphasize two goals: encouraging theparties to negotiate their differences, and incorporating various external costsinto the parties' decisionmaking. 462 This argument brings into focus two pre-mises of a mutual consent theory which are very troubling. The first is the as-sumption that a couple undergoing intense marital conflict will bargain rational-ly, dispassionately, over their gains and losses. But the paradigm of rational

458. The assumption that there would be such effects, and that they would benefitmothers and children, is one of Professor Becker's reasons to recommend change in thegrounds for divorce. See text accompanying notes 119-24. These writers, by the way, donot offer theoretical or empirical proof that the change they propose would have theseeffects. And, once the distributive effects of the new system are recognized, we mightpredict that the law would begin to affect marital bargaining. Would it make men lesswilling to enter into traditionally structured marriages?

459. See text accompanying note 34.460. What Peter Linzer has in mind when he challenges the "amoral" tradition of con-

tract remedies is an efficiency problem, which results from the difficulty of incorporatingnonmarket or subjective values into the usual damages formulas. Linzer, 81 Colum L Revat 112-17 (cited in note 156).

461. The domestic violence problem most dramatically illustrates this point. See textaccompanying notes 168-76.

462. See note 155.

[2:517

Economics and the Problem of Divorce 595

behavior on which a bargaining theory depends may well be less applicable todivorce behavior than any other arena of human experience.46 It is particularlyremarkable to expect that a husband and wife would engage in this type ofnegotiation and conclude that it is better to remain married on efficiencygrounds.4" It is also remarkable to assume that a husband and wife goingthrough this type of conflict will not attempt to negotiate their differenceswithout the spur of a mutual consent divorce rule.

The second premise of a mutual consent theory is that couples can determineall the respective costs and benefits of marriage and divorce, and that they havesome means of valuation which is relatively accurate and which allows them tomeasure and compare precisely those subjective, nonmarket goods which are notwell handled by ordinary liability rules. How would any of us weigh the griefand humiliation of an unwanted divorce against the pain of an unhappy mar-riage? Accepting that these costs of divorce are significant, do we really believethat they can be compensated in money?465 And if monetary compensation isnot feasible, why would it be rational to design divorce rules around the pretensethat they will achieve compensation?

We all know that it is beyond the power of law to command the feeling andexpression of love. There is no legal or economic expertise that can reveal orremedy the causes or cures for marital unhappiness. The argument to deny accessto divorce is an argument which does not respect the importance of marriage.The argument that divorce remedies should attempt to compensate the manypersonal and emotional losses that go along with the failure of love brings lawinto a sphere where it does not belong.

With a contract enforcement approach to divorce, the question of moral andeconomic claims remains important. The entitlements of marriage and divorcecan reflect important moral content, as the norms of fault-based divorce demon-strate. Financial obligations to a spouse or child after divorce also reflect a moralobligation, which is not merely a matter of internalizing the costs of divorce.Where nonfinancial losses are concerned, however, a system of liability rules also

463. Gary Becker has advocated expanding economic models of behavior to includeother, less admirable motivations, including altruism, loyalty, spite, masochism, love,obligation, guilt, duty, and anger. See Becker, 101 J Pol Econ at 386, 398, 400, and 401-02 (cited in note 1).

464. But see, for example, Parkman, No-Fault Divorce at 139 (cited in note 6); Zelder,16 Harv J L & Pub Pol at 247 (cited in note 5). Unlike the commercial situation,negotiations with a spouse over breach are a particularly costly type of conversation. Thisrelates to Lloyd Cohen's point about destruction of quasi-rents. See text accompanyingnotes 163 and 219-23. It is far less destructive to the marriage to seek this kind ofinformation from a stranger. As a lawyer in practice I regularly consulted with clients whowere considering divorce and looking for information about the legal and financial conse-quences they could expect.

465. See note 109 and text accompanying notes 381-82. Presumably, it would take agreat deal of money to compensate these costs. This suggests either that only the verywealthy would be able to divorce, or that some spouses would get only token compensa-tion.

1995]

596 Roundtable

faces the immense problem of constructing entitlements that adequately addressthe moral dimensions of love.

Ultimately, it is much more important to focus divorce remedies on economicclaims. Economic claims are only one category within a larger moral uni-verse,466 but they have a unique character in our popular and legal culture.And, while economics cannot define the entire realm of justice principles, we asa society are more likely to find consensus on efficiency and private orderingnorms than on many other moral questions in the context of family life. Lawand economics have an important role to play in developing sharing norms formarriage and defining the financial responsibilities of parenthood. On the otherhand, law and economics are not particularly helpful in addressing the nature oflove and the problems of commitment in marriage.

IV. Conclusion

This Article has considered normative economic arguments about theproblem of divorce from the perspective of family law and policy. The economictreatments of divorce share a concern to improve the efficiency of our presentpractices of marriage and divorce, and a belief that changes in the present "no-fault" law governing divorce can help achieve this goal. Despite this shared goal,economic theories of divorce law present two distinct approaches. Each reflectsdifferent understandings of efficiency norms and different concerns aboutmarriage and divorce. These different visions lead to contradictory policyproposals, which I have characterized here as "mutual consent" and "enforce-ment" models of divorce.

In the first model, dissolution of marriage is depicted as a transaction inmarital status, which can be made more efficient if husband and wife arerequired to bargain over the terms for their divorce. I describe this model as anapplication of Pareto efficiency norms, and a sort of free market ideal that seeks,as far as possible, to remove law from the processes of family life. The secondmodel presents divorce as one aspect of the larger process of marriage, and lawas part of the framework within which individuals establish and maintainhousehold relationships. In this vision, Kaldor-Hicks efficiency norms are appliedin a search for legal rules that allow parties freedom to optimize the terms forthose relationships.

My criticism of these economic answers to the problem of divorce has twoaspects. First, although both the mutual consent and enforcement models arepresented and defended on efficiency grounds, the choice between the tworequires a choice between different visions of efficiency. Moreover, each solutiondepends on criteria that are not based on efficiency. Thus, the mutual consentmodel requires an allocation of the initial property right in marriage, and the en-

466. Law and economics theorists go to considerable effort to establish that principlesof efficiency and wealth maximization have a strong moral grounding. Posner, EconomicAnalysis of Law at 261-66 (cited in note 2); Posner, The Economics of Justice at 88-115(cited in note 29); Trebilcock, The Limits of Freedom of Contract (cited in note 60).

[2:517

Economics and the Problem of Divorce 597

forcement model requires a framework of liability rules, which may draw oneconomic insights but which also requires other normative judgments. In theseissues lies a large part of the work of family law; by failing to address them, theeconomic analysis remains fatally incomplete.

The second failing of both sets of reform proposals is that they never bridgethe gap between economic theory and the realities of family dissolution. Al-though economics provides a vocabulary for describing many of the factors thatcomplicate these models in the real world-transaction costs, external effects,strategic behavior-these have not been seriously considered. A careful analysisof these problems highlights both the hard choices that policy reform effortsmust confront and the enormous practical constraints on any system of familylaw.

In assessing the contributions that normative economics has made to familylaw, the most interesting question is whether efficiency norms provide a usefulframework for addressing the important policy issues in marriage and divorce.To what extent is the project of family life, and its role in our society, a matterof rational choice, of allocating limited means to unlimited ends?467

These questions are much more difficult to answer in the face of the appar-ent dissensus in contemporary society on norms of family life. In an earlier day,there was at least a greater coherence to the goals and policies of family law,making economic analysis of law and legal policy recommendations far easierprojects.16

' The challenge for family law in these times is much more complex,in all aspects of its work. Given the often widely divergent norms and valuesexpressed in contemporary debate, it is difficult to imagine an efficiency analysiscomprehensive enough to resolve the important policy issues successfully.

467. See Coase, 7 J Legal Stud at 207-08 (cited in note 10) (questioning definition ofeconomic expertise as consisting in the study of "human choice").

468. Compare the discussion in Posner, Economic Analysis of Law at 139-48-(cited innote 2), with Becker, A Treatise on the Family at 350-61 (cited in note 3).

1995]